List of projects

Project information

    Project: Pinehills Logistics

    Team Members: Brian Meehan, Serena Lin, Nolan McDonald
    Keywords: Web application, Azure, User Interface, React JS, SQL
    Abstract: We propose a modern web application to improve the golf bag storage room workflow at the Pinehills Golf Club in Plymouth, Massachusetts. This system allows employees to manage golf bag storage locations, search through existing golf bag locations, and save timestamp records of golf bag removals and returns. The application also alerts employees to retrieve the corresponding bags for members whose tee times are approaching. This project enables Pinehills to handle their daily logistics in a way that is more fault-tolerant, user friendly, and cost-effective than their current solution. Pinehills Logistics is an Azure cloud-hosted web application with a user interface supported by a backend server and database. The frontend is built with React JS, where users can manage the golf bag storage room assignments as well as see the list of golf bags that need to be retrieved for upcoming tee times. The backend consists of an ASP.NET Entity Framework Core system that helps users manage the golf storage room data as well as retrieve tee sheet information. Application data is stored and retrieved from a cloud-based Microsoft SQL Server database.
    Poster: View poster
    YouTube video: YouTube link

    Project: Shepherd

    Team Members: Jordan Blackadar, Ryan Lawton
    Keywords:
    Abstract: The Shepherd software package was created by a team of two undergraduate Computer Science students as a senior design project at Wentworth Institute of Technology. Shepherd aims to simplify the process of distributed resource monitoring in a transparent and simple manner. Shepherd includes software to run on both client nodes and a server computer, which will allow a systems administrator to monitor the health of all computers on the network. After a convenient server setup process, the user simply runs the node software on each computer they wish to monitor. The server software continually monitors these incoming data, such as CPU and RAM usage, and reports anomalous values to the administrator. Leveraging modern Python software libraries, Shepherd involves the use of efficient data structures, multiple system monitoring libraries, three ZeroMQ network architecture patterns, a variety of database tables, and dynamic plotting solutions. Once properly configured, this software allows clients to monitor network performance from a single interface and notifications, rather than active checking or retroactive diagnostics. Shepherd is made available free under the MIT License as a monitoring tool or basis for other software projects.
    Poster: View poster
    YouTube video: YouTube link

    Project: To The Moon!

    Team Members: Martin King, Sancho Rodriguez Corrales
    Keywords: Sentiment Analysis, Time Series, Cryptocurrency, Twitter API, NLTK
    Abstract: This project uses a piecewise linear time-series model that attempts to predict the value of Bitcoin and Ethereum by using closing price and weighted sentiment scores. The Twitter API is used in order to extract tweets from each one of the past seven days with the keyword “Bitcoin” or “Ethereum” with other cryptocurrencies planned for the future. After extracting this data, NLTK’s sentiment analysis’ compound score was used to estimate whether each tweets’ influence was positive, negative, or neutral. From the sentiment score, a normalized sentiment score was created which is influenced based on the number of retweets. A sum of all these sentiment scores was then computed for each day and included as a single value in one of the columns used by the time-series model. This column was used to conclude if tweets’ sentiment score combined with closing price can improve the performance of the time-series model predicting cryptocurrencies’ price. Finally, the project aims to answer whether this model can be used as a valuable tool to predict whether the closing price is going to increase or decrease from one day to the next. We compared modeling the closing price with and without the sentiment values to judge if they were valuable in predicting the next days’ closing price. Our findings showed that including the sentiment value scores from Twitter along with the closing price data resulted in a more accurate prediction of next day closing price which we verified using r-squared values.
    Poster: View poster
    YouTube video: YouTube link

    Project: Cybersecurity Bill of Material Maker

    Team Members: Jake Quilty
    Keywords: Security, open- source, dependency, GitHub
    Abstract: Until recently, open- source software was mostly created and maintained by communities of individuals that wanted to contribute to the free software movement. In the last few years there has been a massive push by corporations to offer a free open- source version of their paid applications. With this push towards corporate maintained open- source, a new cybersecurity threat has emerged: Supply Chain Attacks. The Cybersecurity Bill of Materials (CBOM) is a new concept that stemmed from this issue, and it is basically a list of all dependencies in a project— allowing security teams to know exactly what version of a dependency is being used where. For this project, I took this a step further and created an application that allows security teams to monitor the dependencies that are being used in their projects across an entire GitHub Organization. The purpose of this project is to make it easier for teams to identify which projects contain a vulnerable dependency, as soon as the CVE is announced.
    Poster: View poster
    YouTube video: YouTube link

    Project: LeopardCents

    Team Members: Maxwell Rankins, Corbin Hakimian
    Keywords:
    Abstract: How can I file my taxes? How does one go about buying a new home? How can one pay off their student loans? When it comes to entering the professional world, college students can often be overwhelmed by these types of financial topics that haven't been necessarily covered in many academic settings. LeopardCents is a web tool that unifies all the information that a student may need to know, so one can learn information about their financial decisions and how to start making investments towards their future. It works in tangent with the Student Service Center as a hub for teaching students and faculty about various financial topics, provides a handful of tools to help keep track of finances and allows student to meet with financial advisors to aid in making the best choices for them. Using a combination of VueJS, HTML/CSS and AWS Amplify, LeopardCents provides students with the tools they need to manage their finances during and after graduation.
    Poster: View poster
    YouTube video: YouTube link

    Project: PneumoNet

    Team Members: Dominic Amaral, Robert Martini, Leon Chen
    Keywords: Machine Learning, Transfer Learning, F1 Score
    Abstract: The purpose of PneumoNet is to utilize machine learning to create a model that would be able to determine if chest X-rays had signs of pneumonia. The assistance of machine learning in the medical field aims to reduce doctors' workload and decide which patients need urgent care. Using the model, doctors would spend more of their time on the patients that need urgent care and hopefully save lives. We utilized a dataset containing 10,192 X-rays of normal healthy lungs and 1,345 X-rays of viral pneumonia patients. The ratio between normal X-rays and X- rays of patients with viral pneumonia in the dataset is 1:10 making the dataset very unbalanced. The transfer learning technique was employed to save time on training the model and increase the accuracy with better feature extraction. ResNet50, VG166, and MobileNetV2 were used as base models and compared to see which provided the best accuracy. MobileNetV2, combined with additional layers, proved the highest accuracy. The dataset used to train the model was unbalanced, but the resulting model had an accuracy of 97.46%, with an F1 score of 0.89. In comparison, the average radiologist only has about an 80% accuracy rate when looking at pneumonia lung X-rays. The model is not designed to make a diagnosis of the patient but to only assist doctors in making a diagnosis.
    Poster: View poster
    YouTube video: YouTube link

    Project: KapKeep

    Team Members: Ben Clermont, Nick Nordstrom, Kori Painchaud
    Keywords: body detection, Flask, occupancy, web-based
    Abstract: In the midst of a global pandemic, businesses need to reinvent their technology to conform to the heightened needs of society. Some businesses had this prior to Covid-19, but almost all businesses implement this now: maximum store occupancy. While many buildings have a maximum capacity according to fire code, many stores’ maximum capacities have been decreased due to the pandemic. This means that all buildings now need to keep an active eye on their capacity. Keeping track of a store’s occupancy can be a hassle and a financial burden. It’s unrealistic for a small business to hire an employee to count people as they enter and exit the store. KapKeep fixes this by allowing small businesses to setup cameras at entrances and exits of their store, which will keep a counter of people currently in the store. These cameras use a custom implementation of a histogram of oriented gradients to detect persons in the camera feeds, as well as relay the direct video feed to a website for the owner’s usage. The proposed system employs a single Raspberry Pi camera module and a Raspberry Pi 4B to monitor each entrance and exit. The frames are transferred to a server that monitors the building capacity from each said node. Testing on the Human Detection dataset found our model to be 82% accurate, compared to a popular prebuilt HOG model with 85% accuracy.
    Poster: View poster
    YouTube video: YouTube link

    Project: Spot

    Team Members: Dawn Ramsey and Nicholas Scalese
    Keywords: GPS location, Card stack, Places API, mySQL, Android
    Abstract: Whether you’re moving to a new city, going on vacation, or are simply bored at home, it can be difficult to find places nearby that would be fun or interesting to visit. Spot is an Android app designed to make this process of deciding what to do simpler. Locations are presented to users in a card stack format. These locations can be restaurants, nature trails, amusement parks, and more. Users can swipe right (like) on a location to add it to their personal list of saved locations or left (dislike) to remove it from the stack. This project was worked on by two undergraduate students with the goal of bettering our own knowledge of third-party APIs, backend database development, and UI/UX creation. The resulting Android application performs the main functionalities we set out to achieve. Users are presented with places around their GPS location, can add locations to their list, and can check their list later. There are no plans to release the application on the Play Store in its current state, but this may change as the application continues to evolve.
    Poster: View poster
    YouTube video: YouTube link

    Project: Songworm

    Team Members: William Bernier, Shane Lopez, Nathan Moore
    Keywords: Music streaming service, Spotify, recommendation algorithm.
    Abstract: Finding new songs on the major music streaming services can be a time-consuming task. Oftentimes users search for hours and still find nothing that piques their interest. Spotify has a feature to recommend songs, but it only exposes users to a small number of artists and genres. This is why we propose Songworm, an online database of information about songs and artists that allows the users to rate, comment, and view information about the songs and artists they love. Songworm’s ratings goes beyond the traditional recommendation algorithms used by streaming services by allowing users to rate different aspects of the song instead of a simple like and dislike. This more detailed approach gives a better metric for recommending similar songs to user.
    Poster: View poster
    YouTube video: YouTube link

    Project: StatLeague

    Team Members: Edwin Armas, John Jim
    Keywords: Host Events, Check Schedule, Check Stats, Join Events
    Abstract: StatLeague is a website created to host leagues for sports e-sports teams organizations, or even personal use. Our website allows team captains, players, or organization leaders to be able to access all the information under one place such as, viewing the season schedule, game days, contact information, as well as personal stats: which would be different for every player depending on the game they play as, basketball players having their own unique stats compared to those who play Valorant. Depending on the role, users will be able to create or view events for their league.
    Poster: View poster
    YouTube video: YouTube link

    Project: Accessibot

    Team Members: John Khuc, Xi Xi Lin, Luke Wooten
    Keywords: Discord, Discord Bot, Accessibility, Text-To- Speech, Speech-To-Text
    Abstract: Discord has quickly become one of the most used VoIP and instant messaging applications used by tens of millions around the world. However, discord has not yet been able to offer features that promote enough accessibility for a wide range of its users. Accessibot aims to be an in-app discord bot solution to address accessibility issues targeting those with hearing, speech, or visual impairments. Features include text-to-speech for voice connection, speech-to-text, custom fonts, and easier channel navigation via text commands. These features were accomplished using discord’s api, discord.py, discord.js, ffmpeg, gTTS, and wit.ai. Accessibot and its website, where anyone can invite the bot to their discord server at no cost, are both web hosted on heroku, a cloud service platform. We have tested this personally with our friends on our own discord servers and it was great addition! Through accessibot, we hope that discord will become a more inclusive service for all so that nobody feels limited in their ability to interact with others
    Poster: View poster
    YouTube video: YouTube link

    Project: Infernal Expansion

    Team Members: Spencer Janse, Samuel Ashby, Saurav Dawesar
    Keywords: video games, terraria, modding, AI
    Abstract: The Infernal Expansion is a mod for the sandbox game, Terraria aimed at adding post-game features to the game. This was intended to add to the replay value and challenge for players who were looking for more content to enjoy upon completing the game, which would be done by adding several new enemies, new weapons, and even a new boss. We were unfortunately unable to add the boss, as we determined this would take too much work for the time allotted. We also wanted original sprite work but didn’t have time for this either and instead “reskinned” existing enemies to save on time. We were, however, able to add 5 functional new enemies, 13 functional new tools and weapons, and a new resource for crafting the new items. All enemies spawn in the correct location and the new weapons and the tools are better than the previous tiers. While this is not the full extent of what we had planned, we may still add these additional features in the future!
    Poster: View poster
    YouTube video: YouTube link

    Project: Express-EMS

    Team Members: Zach Sonner, Jason Fisch, Michael Depietro
    Keywords: EMT, EMS, Emergency, Call Transcription, Mobile Application
    Abstract: Our application Express-EMS aims to help those who work in the EMT/EMS professions. The application will be used in response to emergency calls. Such uses include call location, call transcription, and call organization. We aim to make medical emergencies responses more efficient to save lives.
    Poster: View poster
    YouTube video: YouTube link

    Project: DirectPath

    Team Members: Joey Christakis, Giovanni Rico
    Keywords: IPS, mobile app, trilateration, fingerprinting, college students
    Abstract: On average 15 million new college students yearly step foot on a new campus eat year having to navigate and learn their way around a tough and sometimes confusing buildings. Currently there are applications such as Waze or google maps which can get users to a building but lack the ability to help guide you to a specify room. Our mobile app hopes to solve this problem. DirectPath includes an Indoor Positioning System (IPS) that helps to solve this problem by allowing a user to be tracked and positioned indoors. There are normally two choices for IPS Bluetooth tracking or WIFI tracking. In our case to lower the cost and difficulty we looked towards Wentworth’s WIFI infrastructure and made use of that instead. We used the method of trilateration and fingerprinting to get an accurate enough measurement (1-2 meters) of the user’s location. Our fingerprint data helped to mark known locations that we could later reference to help with accuracy. The trilateration method helps to get the users' position within 2-3 meters. By offering not only outdoor but also indoor positioning we can help students relieve their stress and navigate college campuses with ease.
    Poster: View poster
    YouTube video: YouTube link

    Project: Skate Shoe Mate

    Team Members: Alton DeLuca, Randy Ryan, Kyle McDevitt
    Keywords: Skateboard, web-app, exchange, skate shoe mate
    Abstract: Skateboarding is a unique sport, in that the very act of trying a skate trick is destructive to both the skateboard and the shoe performing the trick. The trick is performed with a subtle flick of the front foot, which tears away at layers of shoe fabric. This can result in skate shoes very quickly becoming destroyed and unusable. There are already many products for repairing skate shoes already on the market, and skaters will always be replacing/buying new shoes. Skate Shoe Mate is a platform designed directly towards the skate community. All skaters can relate to having a pair of skate shoes where one of the shoes is in good condition, whereas the other gets beat up over time. This platform is to be used as a retail spot for this type of product. Skaters can come to our website and exchange their beat-up pair of shoes with another skate in the community to get full usage out of skate shoes. Most skaters do not necessarily care about their shoes, so this platform is an option for many and can be a very cost friendly way of approaching skateboarding. Users can post their listings for others to view in the store. Users can also rate each other and warn other buys of a “bad” seller. The main idea of the platform is for users to message one another in hopes of coming to an agreement and exchange their shoes for free
    Poster: View poster
    YouTube video: YouTube link

    Project: Fisherman’s Guide 

    Team Members: Lee Gobin and Taymour Mansi 
    Keywords: fishing, angling, knowledge base, recommendations, community
    Abstract: More people are rediscovering the great outdoors now more than ever. Fishing is an ageless activity that gets people outside and on the water. It is quite easy to fish, but it is not always so easy to catch a fish. Many people rely on relatives or friends that have fishing experience to show them the ropes, but we at The Fisherman’s Guide understand that not everyone is fortunate enough to have a person like that in their lives. That is where our project comes in, The Fisherman’s Guide gives new and experienced anglers years of knowledge in one convenient place. With on the water lure recommendations that let anglers know what to tie on based off the current season, as well as day to day sky and water conditions you will always know that you have the right lure tied on. When anglers are off the water there is a vast knowledge base of articles aimed at improving general fishing knowledge at their fingertips to make them a smarter angler the next time, they find themselves on the water. We also strive to create a welcoming community for anglers from all levels of society to share their knowledge and fishing journey with others. The Fisherman’s Guide truly is a one stop destination for all your fishing needs
    Poster: View poster
    YouTube video: YouTube link

    Project: Energy Consumption Widget

    Team Members: Keegan Whelan, Justin Mann
    Keywords: Data, energy usage, graphs, web development
    Abstract: Every report should have an abstract and it may be included on our website. The abstract summarizes the report using between 100 and 250 words in a single paragraph and is self-contained. The purpose of an abstract is so that a reader can quickly get an idea of the main points and contributions of the project. The author often writes the abstract last. This is because it is only after the body of the report is written that the writer gains a good perspective of the contents and results of the report. All reports submitted to me should follow the format shown in this template. Note that the purpose of the abstract is not to provide an elaborate introduction to the problem, or a motivation for solving the problem, or a discussion of related work. Those analyses belong to the report itself. The purpose of the abstract is to summarize the work done and the results. For many papers, the abstract may be the only part of your paper that anyone reads. Teaming up with another group, the basis of our project was to reduce Wentworth's carbon footprint. Our group’s goal was to create a software solution. Our software solution is a single page application that displays all the residence halls' electricity usage in diverse ways so that students can easily view it and be conscious of it. This was achieved using Hatch Data, a service that Wentworth uses. The whole idea is to have everyone more conscious of the energy usage on campus
    Poster: View poster
    YouTube video: YouTube link

    Project: Adaptive Quiz App

    Team Members: Angel LaVoie, Suman Panta, Sankalpa Kattel
    Keywords: Education, Web App, Adaptive Learning
    Abstract: With this project we accomplished the creation of a Web Application service that serves as a Quiz site for users to test their knowledge and be awarded with harder questions the better they score throughout the various times they access the quiz. The project presented here is a baseline for a larger and broader idea of an AI Driven personalization quiz website that would generate topics and questions for each specific user based on how they answer ones previously for full personal learning experience, this project was used to provide a steppingstone into that idea
    Poster: View poster
    YouTube video: YouTube link

    Project: LeopardBot

    Team Members: Kyle Ferreira, Tyler Kelly, Zack Rosa
    Keywords: Chatbot, Artificial intelligence, IBM Watson Assistant, IBM Watson Speech-to-Text, Conversation
    Abstract: Conversational Artificial Intelligence (AI) is becoming a necessity for large institutions supplying support and information to users. IBM Watson Assistant has become an upcoming disruptive technology, being used globally by 70% of banking institutions. Using IBM Watson Assistant, we can apply machine learning practices to create conversations with our users by using Natural Langue Processing. Leopardbot is a conversational AI chatbot that consults students about life at Wentworth. Conversations are developed using intents and dialogues that the chatbot references when conversing with users. Leopardbot can also be integrated into third party applications including Slack, Facebook, and WhatsApp which allows the Leopardbot to reach a wider audience and provide multiple options for accessibility.
    Poster: View poster
    YouTube video: YouTube link

    Project: Galaga 3D

    Team Members: Nick Quadros, Jon Macone, Garvin Gerard
    Keywords: Gaming, C# Scripting, AI, Arcade
    Abstract: The gaming industry is a constantly expanding market full of opportunities to find success. Single player games have remained popular throughout the years and the evolution of console and PC gaming has allowed game developers to push new boundaries. Our team was really interested in creating a game for our final project, so we decided to bring back an old arcade classic and put our own spin on it. Galaga 3D aims to add a new perspective to the old arcade game from 1981. We were excited to explore new concepts in video game design as well as use the knowledge we have already obtained to create Galaga 3D.
    Poster: View poster
    YouTube video: YouTube link

    Project: Game Music AI

    Team Members: Seamus Nauton, Adam Dzwonkowski , James Ferrarelli
    Keywords: DAW, MIDI, AI: GUI, GODOT
    Abstract: Our project is oriented around music and to accomplish this we decided to divide our project into three parts, AI, the game, and music. Working together diligently we have been able to create a game with a soundtrack capable of sensing where and when you are in the game and play some music accordingly.
    Poster: View poster
    YouTube video: YouTube link

    Project: Autonomous Robot Navigation Platform

    Team Members: David Crafts, Jake Sousa, Nathan Robson
    Keywords: Path Planning, Collision Avoidance, Multi-robot Navigation.
    Abstract: With the ever-increasing demand for consumer goods, companies are pivoting towards semi or fully autonomous robotic warehouses to efficiently fulfill orders. Large companies such as Amazon and Walmart have become increasingly reliant on expensive and proprietary autonomous robotic warehouse solutions. The demand from small to medium-sized companies for similar solutions has increased dramatically. However, these companies don't often have the capital to afford such expensive solutions for their smaller scale warehouses. Thus, we have created a novel, open-source, solution using modern LiDAR sensor technology and advanced algorithms to allow robots to navigate warehouse environments both safely and reliably. The novel combination of sensor data and open-source algorithms will give small to medium-sized companies access to a robust robotics navigation platform, without the overhead of an expensive and proprietary solution.
    Poster: View poster
    YouTube video: YouTube link

    Project: Meal Prep App

    Team Members: Patrick Connolly, Colin Henessey, Andy Dicarlo, Elvis De Leon
    Keywords: Path Planning, Collision Avoidance, Multi-robot Navigation. Meal Planning, Fitness, Nutrition, Exercise, Dieting
    Abstract: Many apps, such as MyFitnessPal and LifeSum, exist that allow you to track what foods you are eating throughout the week, but none are able to offer suggestions for what you should eat to meet your goals. Our app would fix this by allowing the user to input the ingredients needed to make their favorite meals and generate meal plans that fit their weekly nutritional goals. Theapp will also allow users to track the inventory of their fridge and will generate a shopping list of what they are missing to make the meals. With both unique features we will be able to offer users a new way to plan and track their nutritional journey.
    Poster: View poster
    YouTube video: YouTube link

    Project: Collector’s Corner

    Team Members: Zachary Johnson, Joshua Moorehead, James Brennan
    Keywords: Collector, Marketplace, Collection
    Abstract: Collecting items of any sort is a hobby and pastime held by any and all people who have a devoted love for a certain or many categories of items. Being able to track and manage your collection of items can be difficult the more you acquire and while there are a variety of online applications to do different forms of management - none are as complete or easy to use as one may like them to be. In this project, we propose Collector’s Corner, an web based application that provides a complete and feature rich online database management system for people’s collections and a marketplace/community. The collection side of the application allows the users to search for their items and add them to one or many collections. Once an item is in a collection the user will be able to easily view and manage their collections and organize them to their needs. While the marketplace side allows users to buy, sell and trade items with other users from items in their collections as well as being able to chat with others and view their collections. By providing both of these services in a singular application Collector’s Corner provides an easy-to-use and complete application for users to manage their collections.
    Poster: View poster
    YouTube video: YouTube link

    Project: EZ Presence

    Team Members: Omar Abou Nassif Mourad, Michael Marino, James Musacchio, Dillon Morse
    Keywords: Self Service Attendance, QR Code, Classroom
    Abstract: The classroom attendance process is antiquated, relying on inefficient methods of execution such as roll call and manual attendance sheets. The status quo is one of inefficiency and as such would benefit from modernization. We propose “EZ Presence”, a presence management platform that streamlines and speeds up the attendance process by enabling self-service attendance recording. It does this through the generation of a specialized QR code at the start of each class which can then be scanned by students; doing this records their attendance for the session. The teacher also has access to an analytics dashboard where they can view statistics and perform other tasks, such as automatic email generation, to aid their attendance records. By making the attendance process self-serviceable, classes benefit from the saved time and instructors are allowed to focus on the content they are teaching even at the start of the class.
    Poster: View poster
    YouTube video: YouTube link

    Project: OpenWIT

    Team Members: Dylan Goldrick, Jim Garrison
    Keywords: Supervised Machine learning, Predictive Analysis, Student Research
    Abstract: Universities everywhere have a financially and ethnically diverse student population which shares the common goal of wanting to succeed in their studies. There exists both mutable and immutable factors that inhibit student achievement that will be measured in this study. Features pertaining to the individual as well as the institute to attempt to gather insights into this matter. We will survey all University students and store, clean, analyze, and visualize the results. This study will generate a model to predict a student’s GPA, as well as the Institute rating by utilizing machine learning algorithms. General statistical analysis will be performed as well to gather insights on the student population. This will all be anonymous data collection.
    Poster: View poster
    YouTube video: YouTube link

    Project: NoteTogether: A Novel Distributed Note Taking Platform with Blockchain

    Team Members: James Kerr, Max Rioux, Tomas Surna, Fabio Marcellus
    Keywords: Video, Annotation, Note Taking, Blockchain, Ethereum, Smart Contract
    Abstract: NoteTogether seeks to bring interactivity to a very crucial aspect of online learning: consuming video media. We developed a hybrid distributed platform that provides a shared space for watching and annotating video media. Analytics are provided to all users to highlight how users are interacting with a video. NoteTogether utilizes Ethereum Blockchain to ensure data security and scalability by offloading data storage and processing requirements to the distributed Ethereum network.
    Poster: View poster
    YouTube video: YouTube link

    Project: 360 Seating

    Team Members: Rodrigo Moscoso, Martin Cheung, Z Yang
    Keywords: 360-view, Insta360, Movie, Ticket, Seat
    Abstract: With more people returning to the theaters, everyone is eager to get the best seats, but sometimes the seat you see is not what you get. Online movie ticket vendors will display available seats in a generic layout, forcing their consumers to guess which seat provides the best viewing experience. In this project, we propose 360 Seating. 360Seating can raise the attractiveness of neglected seats and even show other qualities of a seat such as the amount of legroom, reclinability, and the number of armrests, by allowing the users to interact with a 360 view of their selected seat. Our project provides a complete online movie ticket ordering system that has similar procedures as other ticket vendor applications but also introduces a point of view feature that separates it from existing solutions.
    Poster: View poster
    YouTube video: YouTube link

    Project: Alert App: No More Waste in Your Refrigerator

    Team Members: Giga Apkhaidze, Abdul Kashem
    Keywords: Android mobile app, food, refrigerator
    Abstract: People used to struggle keeping track of food items in the refrigerator such that many of them were waisted. The goal of this project is to create an Android mobile application that can help people get an alert when certain items are expiring or running out from the refrigerator. The designed Android application not only stores information entered by the users and allows them to check, but it also sends alerts to inform the users when certain criteria are met, such as a food item is expiring. The application also allows the user to filter the list for items, as well as updates its information with a simple click. The design of the user interface takes simplicity into consideration. With this design and a back-end database, we successfully created this Android application for the users to keep track of their food items and get alerts to avoid possible waste.
    Poster: View poster
    YouTube video: YouTube link

    Project: OpenUp: Creating a Remote Locking System

    Team Members: Nick Hinchliffe, Christopher Pizani
    Keywords: Raspberry Pi 4, locking system, iOS application, Wi-Fi
    Abstract: Were you ever locked out from your house? If yes, this project is your solution. OpenUp is a remote locking system and safety solution. Users have the ability to lock and unlock their house doors by a simple click on their iPhone. The system design includes an iOS application that communicates with a Flask server hosted on a minicomputer, i.e. Raspberry Pi 4. From there, Python scripts communicate with a motor attached to the minicomputer. This allows the system to securely lock and unlock a user home. We understand how important it is for a person to feel that his home is protected. To help the users interact with the lock, we designed and implemented an iOS application, which has a clean and simple interface with a great user experience. One benefit of creating an iOS application is that Apple has put a lot of emphasis on mobile security that makes the users feel safe. Within the house, the OpenUp locking system uses a bolt lock and a minicomputer. We found bolt locks are the safest locks on the market and the likelihood of an intruder overpowering it is slim to none. In terms of the minicomputer, Raspberry Pi 4 was used over an Arduino because it has a Wi-Fi feature built in. Our final result, OpenUp, successfully demonstrated that our solution help the users lock and unlock the house with their mobile phone.
    Poster: View poster
    YouTube video: YouTube link

    Project: ShareInfo: Aggregate Information for Investment Decisions

    Team Members: Mohamed Oulal, Caleb Franco
    Keywords: Social media, news, data analysis, Android application
    Abstract: The stock market can be daunting to newcomers who don't know where to look for information. With so many different social media and news media outlets, it is impossible to take them all in. Even worse, existing solutions did a bad job displaying relevant information to the user and giving them an indication of how it correlates to the stock’s performance. In this project, we proposed ShareInfo, a new platform that can aggregate information from different sources, to show the users trending stocks in the stock market, as well as how they correlate with social media and news media. The result of the project satisfied the original goal.
    Poster: View poster
    YouTube video: YouTube link

    Project: Proximity Activation

    Team Members: Dylan Knepper, Jeremy Rojas
    Keywords: Smart home, Phillips hue devices, Android app, proximity activation
    Abstract: Smart Home was a goal that was not yet achieved, but there are still some attempts recently. In this project, we proposed to extend existing standard to realize some of our ideas. Proximity Activation is an Android application that allows Phillips hue devices to be remotely activated and deactivated without the need for the user’s input. The application utilizes the location of the user and periodically checks the distance from the user to a predetermined home location. When the user enters or leaves a set radius from the home location, the device associated with that particular radius will activate or deactivate automatically. The toggling of each Phillips hue device is handled with a Raspberry Pi that is located in the same network as the Philips hue devices. Through the use of the Phue Java Library, the Raspberry Pi was used to access and get the state of each device on the network. The Raspberry Pi can activate and deactivate each Philips hue device and send an updated list of devices on the Philips hue network back to the Android application when requested. The final result of the project demonstrated that we could successfully control Phillips hue devices with the built Android application. In the future, we would like to expand the compatibility of the system, such as enabling and adding traditional home appliances into the network.
    Poster: View poster
    YouTube video: YouTube link

    Project: Wave Music Player

    Team Members: Caleb Smith, Angely Santana, Richard Hui
    Keywords: Windows application, music player, user-generated metadata tag
    Abstract: Traditional music player has a limitation to allow the users to customize songs for filtering and search purpose. The goal of this project is to address this identified issue by creating a software music player that allows customization of a user’s music library. This goal would be achieved by adding user-generated metadata tags into the user’s songs and the users can view and filter songs by properties beyond the standard metadata tags. An example use case of this application would be where the user creates a tag category that describes a song’s mood, such as happy or sad, and then allow the users to attach a mood value to the songs in their collection such that the users can filter the collection to find the specific songs sharing the same mood tag. This project was designed as a desktop application running on Windows, and had a support to play music files stored locally in the user’s device. The structure of the project is split between the front-end UI components and a back-end database. The front-end and back-end components connect with each other to create a library functionality of the application. In current state, the project is still in an early stage, but many of the core features and functionalities from the original proposal have been implemented in a certain level. There are still many additions and improvements can be made. In the future, we would like to improve established features to make them more streamlined and expand the system with more ideas.
    Poster: View poster
    YouTube video: YouTube link

    Project: MedCare4Home: A Medical System for Family

    Team Members: Mengting Wang, Yen Le
    Keywords: medical system, family, Raspberry Pi, mobile devices
    Abstract: Managing healthcare data is essential for everyone, but it can be challenging and time-consuming. Although there existed several Web and mobile applications for personal medical record management, there were few home-based, family-oriented solution. In this project, we proposed to design an easy-to-use solution for a family. We designed and implemented an innovative Home Medical Care System, MedCare4Home, by deploying a MERN stack Web application to a minicomputer, Raspberry Pi 4, and enabling a local network to connect different devices at home. The users can access the Web application using either a desktop, tablet, mobile phone, or simply a 7-inch LCD touchscreen. MedCare4Home not only allows the users to keep track of appointments and organize healthcare documents, but it can also help set up medication reminders and provide self-report interface for symptoms collection for family members easily and securely in one place. The final prototype system demonstrated that we achieved the original goal and beyond. We expect to extend our current hardware design and software development to support physiological data collection with external sensors in the future.
    Poster: View poster
    YouTube video: YouTube link

    Project: BotsideP2P

    Team Members: Adam Beauchaine, Orion Collins, Lukas Van Schaik
    Keywords: P2P, Botnet, Testbed, Kademlia, DDoS attacks
    Abstract: The number of botnet attacks has been rapidly increasing in recent years, so much so that the threat of botnets constitutes a major security consideration for institutions concerned with the prevention of cybercrime. Among the most severe of these concerns are those regarding peer-to-peer (P2P) botnet attacks. These botnets present difficulties in local detection, as well as direct challenges to traditional network security practices due to their decentralized nature. We propose a testbed specialized in the deployment of a P2P botnet with built-in network and endpoint monitoring so that researchers and industry professionals may gain a better understanding of these botnets in a lab environment. The botnet we have deployed implements a distributed hash table to provide bot to bot encryption using the Kademlia. And the Asyncio network framework is used for process handling. All components are implemented via Python. Emphasis is placed on documenting both the structure of the botnet deployed as well as the logging procedures implemented, and flow data and endpoint logs are included in results.
    Poster: View poster
    YouTube video: YouTube link

    Project: Medella

    Team Members: Danielle Shackley, Naishvi Patel, Alyza Diaz Rodriguez
    Keywords: Medical Info Repository, Accessibility, Patient Resource
    Abstract: According to the CDC, six in ten Americans live with at least one chronic illness. Heart disease, cancer, and diabetes are only a few of the many chronic illnesses that are the leading cause of death and disability in the US. Millions of patients are in need of easy to access and reliable information to help them manage their treatment. In this project we propose a website containing reliably sourced medical information where patients can create accounts and save articles available in 64 languages and listen to its audio version.
    Poster: View poster
    YouTube video: YouTube link

    Project: Drink Match

    Team Members: Ryan Ng, Ethan Lopez, Kevin Bruno
    Keywords: Dating App, Matching, Drink Preference
    Abstract: Methods of meeting new people has in a sense, changed from meeting in person to meeting online, especially due to the COVID19 pandemic. We can see this through the abundance of meeting and dating apps available on any single smartphone or web app alone. Most of these apps match specifically based on gender/ sex and some randomly entered interests. We wanted to try and put a spin on this idea by matching people solely based on the drinks they like. Quite literally any drink, anything from normal tap water to vodka cocktails. In this project, we propose a smart phone app that can match people simply on the drinks they like plus some individual preferences. Providing a unique and new way to reach out and meet new people based on a simple line of interest and taste.
    Poster: View poster
    YouTube video: YouTube link

    Project: VICTR

    Team Members: Marcus Kwong, Alyssa Lohr, John Thebarge
    Keywords: Virtualized Infrastructure, Virtual Machines, Testbed Platform
    Abstract: Virtualization software and platforms have become increasingly popular over the past few decades. Virtualization allows companies and universities to save money and resources while maximizing the potential of their equipment to host all the services and functionalities they need. Virtualized Infrastructure with Cyber and Testbed Resources (VICTR) is a ESXi based virtualization platform that provides Wentworth Institute of Technology (WIT) students and faculty with access to a virtual machine environment that is hosted on WIT’s internal network resources. By using VICTR, students and faculty can access multiple virtual machines to conduct labs, testing, and research without sacrificing the computing resources on their personal computers.
    Poster: View poster
    YouTube video: YouTube link

    Project: HearthFoundry

    Team Members: Marcus Wong, Gabe Ehrlich, Kevin Liu
    Keywords: Searchable Database, Online Card Game Tool, Text-based filtering
    Abstract: HearthFoundry is a fully searchable database tool for Hearthstone, an online card game where you collect cards and challenge other players. HearthFoundry contains a database of all the relevant cards along with everything that makes the anatomy of the card, such as its statistics. We found other existing card database tools for the game to be lacking in powerful search features such as text-based filtering. HearthFoundry has user-friendly features like being able to search for any card in the game, view its information, and easily filter for cards based on specific parameters. Our project leverages NextJS’s dynamic routing to keep the URLs clean, and allow navigation without entering the webpage, and MongoDB as the backend database.
    Poster: View poster
    YouTube video: YouTube link

    Project: Out There

    Team Members: Devin Denham, Marcus Pongratz, Tim Reilly
    Keywords: Dating Service App, Matching, Stable Marriage Algorithm (Gale Shapley Algorithm)
    Abstract: Out There is a Web Based Dating service. Out There revolves around a new and unique type of matching algorithm which is what we expect to separate it from the rest of the mundane dating services already in existence. In our quest to create a better model for finding friendly, creative, and active people we implemented a second layer of matching input for users that find a match. We allow users to “Like” or “Dislike” profiles that come across your screen like any other service, but we also ask you to RATE your match on a scale of 1-10. When both parties have rated the match, both parties can be informed of the other’s rating, giving both users a more trustworthy/authentic look into how much the person on the other end is interested.
    Poster: View poster
    YouTube video: YouTube link

    Project: M.M.B.S

    Team Members: George Wood, Juan Kochen, Miles Macchiaroli, Tylor Hughes, Nick Portanova
    Keywords: Sentiment Analysis, Social Media Data Aggregation, Financial Analysis, Meme Market
    Abstract: Modern Media Business Systems, or M.M.B.S, is a financial services project aimed at providing real-time data, notifications and alerts from user created metrics. By analyzing both social media and traditional metrics, M.M.B.S is able to allow users to take advantage of the “meme market”, avoid social media “pump and dumps” , and make long-term smart plays. The mission statement of this project is to make smart trading for anyone, regardless of your financial comprehension, ability to spend hours monitoring social media, or size of your bank account. M.M.B.S utilizes sentiment analysis, social media data aggregation, financial data ingestion and analysis (algorithmic), and various forms of alerts including SMS (Text) and SMTP (Email). In 2021 we have seen Gamestop ($GME), AMC ($AMC) and Dogecoin (DOGE) all become not just household names but life changing meme market opportunities. By using M.M.B.S, a user would not miss out again, and instead would get in, in time to make a hefty profit.
    Poster: View poster
    YouTube video: YouTube link

    Project: HotSpot

    Team Members: Akeylah Roscoe Hunter, Sophia McGlew
    Keywords: Interactive Social Platform, World Travel Experience, Planning Vacation
    Abstract: Social Media is one of the most popular online presence for people around the world. Social Media users tend to post their vacation travels on their accounts, leaving others wondering where they went and what they are doing. Vacation Travel is also in very high demand as the Pandemic continues to slow down. Vacationers tend to use websites like TripAdvisor, Travelocity, Yelp, and Expedia to find places to stay and eat on vacations. However, some existing websites and applications do not have accurate reviews based on different locations. The following applications and websites do not have a one for all platform, that allows users to post their travel endeavors while finding new places to eat, stay, and things to do. In this project, we propose HotSpot, an interactive social website for people to share their travel experiences and to inspire other users who are planning their vacations. Travelers will be able to find specifically what they are looking for when planning their vacation, like restaurants, hotels, activities, and many more.
    Poster: View poster
    YouTube video: YouTube link

    Project: DenCity

    Team Members: David Edwards, Joseph Spanilo
    Keywords: Wireless Localization, Crowd tracking, Data collection
    Abstract: Google Maps has made it easy for people to plan their commute based on how busy a road is. This has also been expanded to help users understand how busy a store may be during the day. However, this feature has not been applied to buildings with much depth. DenCity aims to solve this problem by showing a way that the number of users can be tracked without implementing any new hardware to existing wireless infrastructure and without the user having to participate. DenCity uses the WiFi signals sent out by devices to track the number of users near an access point. As a result, it makes it possible for users to know the number of people within a building or given area.
    Poster: View poster
    YouTube video: YouTube link

    Project: GamingZone

    Team Members: Daniel Johnston, Dexter Chan
    Keywords: Gaming Blogs, User Friendly Design, Livestream
    Abstract: Gaming blogs and websites talk about games and give their own personal review of them. There are a lot of platforms out there that share the latest information about games that can match your interest. However, there are a lot of people that use their own platform and those wanting a review will have to open multiple sites. Gaming Zone offers a way for everyone to give their own reviews about a game. Gaming Zone will offer a user-friendly and stylish website for its users. Not only will people be able to give reviews, but also be able to livestream their favorite games and share it with everyone.
    Poster: View poster
    YouTube video: YouTube link

    Project: Runner’s Training Center (App)

    Team Members: Matthew Bishop, Isabella MacMahon
    Keywords: Android App, Running, Tracking, API, Java
    Abstract: Several studies have shown an enormous wave of new runners worldwide. As a result, a market dedicated to the sport has emerged. Athletic companies have been producing more varieties of cutting edge running footwear, apparel, and software applications. While all market varieties have contributed to the amount people have been able to push themselves; none have contributed quite as much as software. Each new application has bettered runners' abilities to track their runs, find routes, and share their successes with others. Yet, despite the growing number of applications available to runners, there are very few apps dedicated to the selecting and tracking of training plans. A task that, as many runners are well aware, is quite challenging and time consuming. In order to streamline this monotonous and tiresome task, an Android mobile application: Runner’s Training Center, was developed. The mobile app was designed with the intention of assisting runners in selecting and tracking training plans. Additional features include pace calculations, coaching tips, and personalized statistics. Furthermore, the app allows users to peruse through plans, rate their daily runs, share their completed plans with others, and browse cross training resources. To accomplish this, Android Studio, Java, Google Firebase, JSON, PDF Viewer API, along with many more APIs, formulas, and algorithms were used. The resulting application is now available for use on the Google Play store.
    Poster: View poster
    YouTube video: YouTube link

    Project: MyWIT: A Social Media Platform For Wentworth Student

    Team Members: Zhiheng Chang, Yihao Li
    Keywords: Social software, College based
    Abstract: With students across the country attending their classes online rather than in person during the pandemic, we are left to ask how we can make school social again. Currently there is no platform online to help students to solve this problem. Wentworth has switched their online platforms to newer and simpler website called myWentworth. However, this website does not serve much use to their students and fails to benefit their college experience. Its main purpose is to be a campus hub that holds links to all other Wentworth related websites. They have missed an opportunity to launch a web application to make their students’ lives more social again. Our project aims to bring social opportunities back to Wentworth students while fully online. Instead of a collection of campus resources, our project would be more beneficial for the students to have a social media site in difficult times like these. Many students have difficulty staying in touch with and meeting new classmates while not in person. A social media application will certainly bring its students closer together and serve a real purpose. Students will once again be able to share what they are thinking and communicate as if everyone was back together again.
    Poster: View poster
    YouTube video: YouTube link

    Project: A-Trade

    Team Members: Nicholas McCartney, Matthew Stepnowski
    Keywords: Investing, Stock, Cryptocurrency, Technical Analysis
    Abstract: The average interest rate on a savings account is much lower than the rate of inflation. Your money loses value sitting in the bank unlike investing where the average rate is higher than inflation. Today, some people do not know where to get started when it comes to investing. They are either too afraid, or unsure how to invest their money safely. A-Trade is a web application designed to assist traders of all levels whether they are completely new, or experienced. Using A-Trade users can customize popular technical analysis algorithms and compare them to buying and holding a stock. By providing the returns of both, the technical algorithm and the simple buy and hold, users can realize that investing isn’t as complicated as they think.
    Poster: View poster
    YouTube video: YouTube link

    Project: Melodyze–Algorithmically Generated Melodies

    Team Members: Vedavyas Munugoor, Jacob Balsamo, Andrew Canario
    Keywords: Melody, Generation, Intuitive, Songwriting
    Abstract: One of the most common issues facing songwriters and music artists today is writer’s block. Existing music software to produce melodies and background music focus almost entirely on giving a music producer a “blank slate” with many options for what instruments to add. While these applications are very powerful, they can also be very intimidating and give too many options to a musician at once. We propose that Melodyze, an Android platform mobile application that algorithmically generates short melodies, can fix this issue of writer’s block by providing musicians short melodies that they can draw inspiration from. A lack of creativity can be a significant obstacle that musicians face. With Melodyze, we hope to streamline the songwriting process for full-time creators and artists.
    Poster: View poster
    YouTube video: YouTube link

    Project: Safety Violation Identifier with Images

    Team Members: Raymond Ren, Danny Wong, Andrew Zemko
    Keywords: Viola Jones Technique, COVID-19, Social Distancing, Face Detection
    Abstract: In March 2020, the United States, and nations around the globe, suffered an outbreak of a highly contagious virus known as SARS-CoV-2, better known as Covid-19. As a result of this global threat, many different countries, including our own, resorted to a mandate of face mask usage in order to reduce the spread. While this was a successful attempt to stop the disease in its tracks, there remained one important question: How can we monitor this? With our software, this task becomes simple for anyone at the helm. Once a functional webcam or camera is connected to the host device, our software will use what is known as the Viola Jones technique to isolate individuals’ faces and report whether a mask is present in real time. Once this is determined and visualized for the user, they can then decide what to do with that information.
    Poster: View poster
    YouTube video: YouTube link

    Project: HumTech

    Team Members: Paul Sender, George Wishart, Steven Russman, Gavin Fisher
    Keywords: Image Processing, Artificial Intelligence, HVAC systems
    Abstract: HVAC systems provide an easy way for building managers to control the climate in their buildings. Currently, these controls are done manually and as a result there is a lot of wasted energy. I.e. HVAC systems are kept on until they are turned off. HumTech provides a solution to the problem by using AI technology to determine if there is a person present in an image taken by a simple USB camera. If there is a person present, the HVAC fans turn on, if there is no people present. They turn off. This essentially automates the management of these systems saving building owners money on their energy bills. To achieve these results, the HumTech team used a raspberry pi 4, and an open source image processing solution called ImageAI. Every 90 seconds the python script tells the Pi to take a picture using the fswebcam module then sends that picture to the ImageAI software so it can be processed. Upon completion, the python script then goes through every object detected in the image and determines if it is a person or not. If it is, the count of people increases. The system then sends a GPIO signal to pins 0 and 1 that then further the signal to a VAV controller that tells the HVAC fans to turn off or on.
    Poster: View poster
    YouTube video: YouTube link

    Project: Image to Text Translator for Visually Impaired and Blind People

    Team Members: Alisha Rizvi, Caitlin Morales , Jason Cheung
    Keywords: Visually Impaired, Screen Readers, Translator, ImageNet
    Abstract: 8.1 Million people (3.3% of the world population) are blind or visually impaired. There are not many known facilities to help this population with the use of technology, especially the use of the internet. One main facility is screen readers. Screen readers are used to help aid in using the internet by reading aloud everything on the screen but often times they cannot read images unless alt text is already provided for the image either through captioning or html text. Without the functionality of reading and translating images, using the internet becomes a hassle. We propose the Image to Text Translator application. It’s an application whose title speaks for itself. By using ImageNet dataset and a pre-trained model with 92.7% accuracy, the image to text translator application is able to take in an image and subsequently provide a text and audio translation of said image.
    Poster: View poster
    YouTube video: YouTube link

    Project: Parking Spot Counter

    Team Members: Munoz, Avery, Everett, Corey, Duong, Phat
    Keywords: App, Image Recognition, Machine Learning
    Abstract: Wentworth students can potentially face a full parking lot when they arrive on-campus. This can potentially impact their class times and grades. As such, they need a way of knowing whether or not the parking lot will be full when they arrive in order to prosper as students. There is no current way to check for space. In this paper, we put forth Parking Spot Counter, an application that can count the number of spots available at any given time using a Jetson and a full-stack application that takes pictures every 5 seconds and gives real-time data on how many spots are available to students, resulting in their accommodating for certain times by looking at historical data and planning accordingly. The goal of this image processing project is that no student will arrive to the parking lot without a good idea of what they should expect, and to ultimately make the experience of knowing when to arrive at Wentworth easy.
    Poster: View poster
    YouTube video: YouTube link

    Project: Safe To Face -Android Application

    Team Members: Meghan Sicard, Schebania Joseph
    Keywords: application, personal safety, Android Studio, SQLite
    Abstract: Safe To Face is an Android application to be used for personal safety. Studies show that 50% of women feel unsafe walking alone at night, 13% feel unsafe getting a taxi or ride-share by themselves, and 21% feel unsafe going on a first date. Many women when alone or in an uncomfortable situation will call someone on the phone to help ward off unwanted attention. Unfortunately, not everyone has someone they can rely on to answer. This application allows users to interact with “fake” phone calls where the recording on the other end will make it known to those around that they are waiting for them and know where they are. As an additional safety feature, this app also allows users to share their location with all of their emergency contacts at the push of a button. While these features should not replace emergency services in a dire situation, it can provide peace of mind and prevent unsafe interactions with strangers.
    Poster: View poster
    YouTube video: YouTube link

    Project: Overtime Schedule

    Team Members: Sean Crowley, Jake Dickinson
    Keywords: Schedule, shift coverage, excel, management
    Abstract: Many businesses today are still using paper schedules for their operations. Each individual employee must memorize their schedule that is written down in the shop or take a picture on their phone to know their schedule each week. On top of this, if coverage is needed, they must identify who is available to work and contact each person themselves. From our personal experiences, these things can lead to a lot of headaches for employers and employees when it comes to remembering shifts, employees/coworkers not showing up on time, and finding coverage for your shifts. Overtime schedule aims to eliminate all these problems by making the entire process easier for both the employer and the employee. Overtime Schedule provides capabilities for employers to easily add employees/managers and create schedules for their business inside the application or by exporting to excel, as well as empowering employees to easily view their schedule, request time off, and auction their unwanted shifts to coworkers. Overtime schedule is an all-in-one place for businesses to stay organized and make life easier for both employers and employees.
    Poster: View poster
    YouTube video: YouTube link

    Project: Boston Fitness

    Team Members: Alex Brito, Gary Hui, Rana Mohmedsiddik
    Keywords: Android, mobile app development, Java, social fitness, exercise
    Abstract: High obesity rates of adults in the U.S. are growing at an all-time high. A simple solution would be to exercise; however, that is difficult to act upon due to lack of motivation, consistency, or support group. The purpose of this project is to combat these concerns by offering people an outlet to exercise together. Our conjecture is that having another person is a source of encouragement to motivate oneself. Our solution is to create Boston Fitness, an Android application primarily focusing on allowing people to create events to gather and perform any exercise—making it unique to any other social fitness applications.
    Poster: View poster
    YouTube video: YouTube link

    Project: Automatic Joint Space Width Finger Measurement

    Team Members: Tino Cheung
    Keywords: Joint Space Width (JSW), Osteoarthritis (OA), Image Processing, Machine learning, Tuft Medical Center
    Abstract: Radiologist captures x-ray scans of their patient to then be digital images that doctors can review and create medical reports. The medical report is sent back to the radiologist where he/she can explain to the patient on what the diagnosis were. However, a bottleneck occurs when the quantity of x-ray images become overwhelmingly great for the doctor to handle in manual labor. Biomedical research groups have applied their knowledge of image processing and machine learning to help create software that can have the potential to solve this manual labor. In this paper, I propose a method for automating measurement of joint space width (JSW) to x-ray finger joints that are labeled with Kellgren and Lawrence (KL) grading of osteoarthritis (OA). The methods applied for an automatic JSW finger measurement was using machine learning and image processing techniques to better achieve reliability in feature extraction of the joint of each finger (excluding the thumb). This project was made possible with the support of the members in Digital Health Lab associated in Pace University and Wentworth Institute of Technology. The collective data of x-rays and manual JSW measurement data sheets were provided in collaboration with Tuft Medical Center.
    Poster: View poster
    YouTube video: YouTube link