School of Computing and Design
Computer Science Capstone Festival Spring 2021
Text-to-Video Conversion Application
Parker Erickson | Edgar Reyes | Aaron Helman | Emory Schmitt
Individuals and or companies wanting to express their ideas through videos on social media platforms may be unable to create the video content they desire due to limitations. These limitations may include the lack of time, lack of resources and lack of feeling comfortable in front of a camera. Text-to-Video website aims to meet the video creation needs of an individual and or company by converting written text into an engaging video that includes images, audio, and subtitles.
MBARI Video Annotation
Frank Piva | Federico Rubino | Kyle Dilbeck | Jerry Kankelborg
MBARI has collected 30,000+ hours of deep sea video that needs to be annotated. Our tool uses machine learning to render the videos and feed it to their Cthulhu video player and annotation tool. The localization tool (Cthulhu) needed improvements with implementing bounding boxes to be editable and resizable, this is possible now. The Machine Learning community lacks tools to properly render annotations and localizations on video and therefore we have created a second tool called Kassogtha to be a support tool for Cthulhu as well as other video annotation tools.
The Last Jet FightAR
Josue Gonzales | Christiana Libhart | Jose Alfaro | Cathy Hsu
When fighter pilots are piloting their aircraft in the midst of battle, they must contend with both the physical strain of flying as well as the mental pressure of deciding how to position their aircraft to outmaneuver enemy opponents. Incorporating AI into the fighter pilot system would relieve pilots of some of that mental fatigue by having the AI plot the plane’s optimal path instead. This team’s goal was to develop a demonstrational prototype for an AI-Human teaming for guidance aircraft fighter pilots against adversarial aircraft. We created a simulation of the high yo-yo defense, an aircraft maneuver that allows a pilot to position their aircraft behind the enemy plane’s blind spot. Whenever the player pilot deviates from the high yo-yo defense path, the simulation will use guidance systems to direct the pilot back onto the optimal path. Ultimately, we hope this prototype demonstrates how AI-Human teaming can save lives and improve the survival rate of fighter pilots.
Search and Rescue - Game - AR Prototype
Ethan Herndon | Riley Galloway | Natalia Gill | Athena Raya
This capstone project supervised by our client Clay Gruenke is a Search and Rescue prototype that uses Augmented Reality (AR) using the resources available in Unity; a game engine platform. This project sets out to address the difficulties that can arise when conducting search and rescue work i.e., limited human capabilities. The current prototype has the user interface (UI) setup in a way where the rescue worker can navigate and switch through different search methods using buttons. The rescue worker has the ability to track where the grounds searched were already visited using Shaders and a Shader-Graph System in Unity. The ability to communicate between multiple devices and all necessary networking components are currently in the process of being implemented.
Gamification (Talent Acquisition) - 1
Rocco Estrada | Jaclyn Libhart | Julio Aguirre | Nathan Latronica
Capital Insurance Group would like web-based games for talent acquisition purposes. Our goal is to create a web application that will both attract and assess applicants. This assessment will then help CIG determine if the applicant(s) have basic programming skill and knowledge before the interview process begins.
My Cheat Sheet for Voting
Isidro Perez Zepeda | Cody Todd | J. Eduardo Garcia M. | Blayne Suttonwills
To ease the process of voting, My Cheat Sheet for Voting allows users to have all the information they need to make more educated and informed decisions about who or what they can vote for in their upcoming elections in an application directly on their phone. The application allows users to view information on their local voting options and list what they plan to vote for in each election as well as uses login information to save data if the user wants to use multiple devices.
Code Submission System for Energy Usage Testing
Aldrin Dancel Carlos | Eddie Sanchez | Blessing Apkarian | Deborah Meda
The ultimate goal of Green Code is to motivate Computer Science students to learn. There have been studies that show Generation Z, the current generation of college students, i.e. Computer Science students care deeply about Climate Change. We hope to use Climate Change to motivate students to learn and create more efficient programs. Our aim is to create a new measurement of program efficiency like Big O of Time and Space such as Big O of Energy. This project features an interface where a user can upload a file and, within a few minutes, receive an energy report based on the performance of the file. The report is generated by running the program on a remote raspberry pi where energy sensors and monitors will track the energy consumption throughout the program's compilation. The user can view all the files they have previously uploaded with the energy feedback for each upload.
UI Design for CalSPEED Mobile App
Pedro Munoz | Christopher Scott | Alvin Liang | Mustafa Memon
We developed a new UI for California Public Utility Commission’s CalSpeed Android app which allows a user to do multiple tests including mobile data download, upload speeds, the ping etc. We made the app more engaging by adding animated visual cues which inform the user how the speed test is progressing. We also took a minimal approach to the layout of the app to reduce the amount of information displayed on the screen unless the user interacts with the app.
Cluster Estimate Tool
Jesus Anthony Perez Zepeda | Rowan Briggs | Jesus Caballero | Carmelo Hernandez
Annually, vineyards must calculate the estimated yield of grapes in order to know more or less how much they have to sell. The current method has proven to be ineffective due to human biases that occur when the person in charge of the process is out in the field. The estimation is a three-part process: cluster estimation, grape count per cluster, and weight of cluster. Our project focuses on the first step cluster estimation which is the most crucial given that it is responsible for 60% of the overall estimation. It will allow workers to input the size of the vineyard and give the worker points at random to inspect until the desired confidence interval has been reached, helping to reduce estimation errors.
Independent Worker Safety Escalation Software
Quynh Nguyen | Anthony Rubio | Jose Morado | Albert Salas
Web application that includes a dashboard where managers and supervisors can view an employee’s current location and receive notifications if the employee has not moved recently, thus, ensuring the employees’ safety.
Optimized Allocation of Computational Resources
Tracy Riordan | Jonathan Cabrera | Zhaojia Liang
Computer Resources are valuable and finite. They can only run a certain number of programs at a time. If the resources are not used efficiently then downtime will stack and applications that are critical to the system may lose precious time waiting for the system. This project aims to minimize this downtime by ensuring that the application has enough computing power to be done as quickly as possible while minimizing the number of computers necessary.
Course Materials Search Engine
Yvonne Cruz | Ozzie Moreno | Joshua Trup
When a student finds themselves searching for course materials they often skip to google because looking through Ilearn can be time consuming. This is a problem because they never find the genuine lesson taught by the teacher. We created a search engine website that can search through files in a google drive shared by a professor. This will help students find course information such as pdfs or videos much quicker than previously and these materials will be the original lesson given by the professor. Our project has a different page for each course and new materials can be added easily with google drive share permissions.
Ready Pallet to Cool
Richard Ledbetter | Miguel Espitia | Efrain Pamatz | Ben Zhang
Due to weaknesses in the current transportation process of pallets, we have created an android application that is an essential tool for workers. The tools we used for our app were android studio, dynamo database, Google authorization, and ML Kit text recognition. Food Origins noticed room for improvement in the current system as more often than not pallets can be mixed or lost in transit and can lead to payment loss for some workers. To help mitigate this, using the tools previously mentioned, our app allows for the sender to take a photo of a pallet’s label, then using ML Kit recognize the label’s text. Then, that text is stored along with the GPS location of the user into a dynamo database which is also viewable. Finally, the receiver will take a photo of the pallet label again to acknowledge its arrival and their GPS location is stored as well. Overall, we hope this app becomes a valuable tool to help essential workers.
Gamification (Talent Acquisition) - 2
Christian Jimenez | Jonathan Quintero | Nick Tralongo | Geoffrey Woods
Gamification (Talent Acquisition) - 3
Kevin Piffero | Brian Baird | Mytzy Escalante | Rodrigo Espinoza
We have worked with CIG to implement a strategic assessment that will diminish the chance of a non-qualified applicant being able to advance to the next step of the interview process. In order to achieve this, we implemented a few different types of assessments to gauge the applicant’s knowledge of Java by working with CIG to screen new junior engineers. With this knowledge, CIG would then be able to have a quicker turnaround time for the interview process. The applicant will be assessed on their technical skills in an interactive way.
Blockchain Identity or Supply Chain
Marc Alejandro | Jorge Lopez | Charlie Nguyen
There is limited visibility within the medical supply chain when it comes to receiving medications or vaccinations. The process of looking up a patients and their vaccination statuses, especially online, is virtually nonexistent. Our blockchain-based tracker tracks every turn of a patient's vaccination process, from approval to second dose. This will allow employers or any other entities to verify that a person has been fully vaccinated.
DeepSea-AI Visual Event Grid App
Shawn Deppe | Christopher Jimenez | Eduardo Soto Rodriguez | Kevin Guzman
MBARI has been conducting deep-sea research that has resulted in over 27,000 hours of underwater footage. They are currently in the process of digitizing this vast library. In doing so they are looking to machine learning to aid in the annotation process. Our objective is to create a web application that will help the MBARI team with identifying false positives (Incorrect guesses) generated by their machine learning algorithm. Due to the vast amount of underwater footage collected by MBARI, annotating the process is time consuming. Our aim is to create a program that will help with reducing the time required to annotate data from the video.
GDELT Browser-Based Access Tool (GBBAT)
Claudio Vertemara | William White | Liam Cunningham | Denize Ignacio
GDELT is a massive open database that monitors the world’s news media from articles to live news coverage. Over this semester, we have worked on a tool called the GDELT Browser-Based Access Tool (GBBAT). GBBAT aims to make GDELT more accessible, allowing users unfamiliar with SQL to make queries and filter the GDELT database to their needs. In addition, it summarizes the information in a visual format. This is a continuation of previous Capstone groups. For our group, our work focus was twofold: the first was moving GDELT data to a different server, and the other was making GBBAT more user-friendly.
CSUMB Intramurals App
Ryan Beck | Cass Cabrera | Joey Thomas | Jake Fahy
The CSUMB Intramurals app was created as an alternative to IMLeagues, the current application CSUMB uses to organize their intramural sports. We kept the best features of IMLeagues such as creating/joining teams, generating game schedules, and keeping track of scores while also addressing the main issues of the application, such as removing advertisements, creating a more user friendly application interface, and using industry standard software to ensure the application functions as intended.
Anxiety Data Musings - Foresight Health Solutions
Emmanuel Castillo | Victor Cuin | Misael Guijarro | Jacob Ortiz
Anxiety continues to be a huge issue that impacts many adults across the nation. It is estimated that about four in ten adults in the U.S. suffered from increased anxiety during the pandemic. In an effort to help people better understand their personal well-being, we created a web application that can give better insight into their current level of anxiety. Using a machine-learning algorithm, we return a predicted anxiety score based on the user’s answers to our questionnaire with a bit of insight on what that score means. We also provide extra features to registered users. We keep track of a registered user’s past scores, which they can access at any time to see the trajectory of their anxiety over time. These users also receive recommendations that may help calm their anxiety when practiced. We keep track of the effect these suggestions have on users.
Regression Test Analysis Tool & Dashboard
Jason Buckley | Joseph Villegas | Jonathan Welham | Kailun Yang
We developed a data analytics dashboard in the form of a containerized web application using Docker for Capital Insurance Group (CIG). This application aimed to organize and understand the results of their test suites by providing an actionable analysis of each of their suites progress. The dashboard's backend is able to accept, process, and deliver test information from a SQL database via endpoints in our Rest API, and also deliver email notification of tests' progress categorized by app using the Python Flask framework. The frontend is able retrieve, process and visualize test data provided by GET requests to our server using the React framework.
Round Trip Time Analysis & Visualization CalSPEED
Jennifer Lopez Hernandez | Ethan Castro | Jesus Gomez | Eric Guerra
RTT Network is a web-based application that allows the general public to understand broadband round trip time (RTT) trends between four different carriers. We used data provided by CPUC and the plotly library to plot multiple graphs. This way we can explain what these trends mean and allow the general public to make an informed decision when choosing one of the four carriers.
Get Up and Move
Alexis Caasi | Roober Cruz | Bobby Davis | Raul Perez
Whether it is Netflix, video games, or work, people today have many reasons to sit for hours at a time. Studies have found a strong link between sitting for long periods of time on a regular basis and cardiovascular disease. Our team has developed an android app that reminds people to get up and move throughout the day while also educating them on the dangers of sitting for too long. Users are encouraged to get up and follow along with an exercise guided by an animated avatar.
Multi-Purpose Exploitable Server
Alicia Sandoval Garcia | Tessa Rosasco | Timothy Hanneman
Network and computer security are a major concern in today's computing environment. Cybersecurity attacks have the potential to cause extensive monetary damage to an organization. We have created web server software with insecure modules that can be utilized by those involved in cybersecurity to practice their pen-testing skills. Of the insecure modules that the server contains, one is exploitable with command injection through the URL string. The server includes logging to detect potential network intrusions. The user also has the option to start the server with a graphical user interface, or from a shell. With the use of our software: cybersecurity novices can increase their knowledge of common attack methods, and detect network intrusions when utilized as a honey-pot.