Focul
Inspiration
While discussing potential creations for this challenge, we decided to attempt solving a looming problem of our lives– concentration. Due to the common distractions of our everyday lives, many people suffer from a poor work ethic, which leads to procrastination, less sleep, and worse health overall. We wanted to create an app that would be useful for everybody and could be used over and over again. On top of this, we wanted to make our app perfectly suited to each individual user so that each person would find success in our project. The adaptability of the neural network was one of the main emphases for making this app much more personally suited for each person.
What it does
Through simple yet effective GUI, our app is able to record specific landmarks of the user and translate them into a neural network that will be able to detect whether the user is focused or not, helping them through their daily tasks, reminding them when they start to drift away from their tasks at hand. Through customizable on-screen and audible warnings, the user will be sure to stay focused while they utilize the app.
How we built it
We utilized a custom perception in order to create a neural network that would read the landmarks of the user. The neural network will also only utilize landmarks, significant points on the user’s body, such as parts of the eye and mouth, to decide if the user is focused or unfocused. Within the graphical interface, the user can complete all necessary tasks to allow the program to function, including additional sections to allow for incredible customizability. After the neural network is trained based on this data, it will utilize a live video that feeds detected landmarks into the neural network to decide whether users are focused or not. If the neural network detects that the user is unfocused, the user will receive both a visual and an audial warning, to bring the user back on track. This app is packaged in a simplistic and intuitive graphical user interface in order to avoid the distractions of an unnecessarily complex program.
Challenges we ran into
While programming our app, we had problems with optimizing the neural network. We slowly realized the immense amount of tuning needed in order to create accurate neural networks of image-based data, which meant that the project would become much more difficult than anticipated, deviating from our planned schedule and timeline in order to complete this project on time. However, we were able to address this by dedicating more time to the project in order to have the best results possible.
What we learned
Throughout this experience, we were able to learn how to incorporate a complicated back-end with a smooth front-end. We learned how to develop a neural network for our app that required machine learning. Afterwards, we struggled creating a GUI that would allow us to comfortably display our program, but we eventually found success our integration of python's Flutter module.
What's next for Focul
In the coming months, there are many improvements possible for the app. We would transfer the code written for the graphical user interface from Python into a different language such as C# to make it much simpler to integrate functionalities to benefit the user. We would add functionalities like being able to set up different user profiles so that multiple people could access and utilize this app from the same device without needing to re-record data. Along with this, we would also make it much simpler for users to access by uploading it as an app that the public would be able to access and use in their daily lives.
Log in or sign up for Devpost to join the conversation.