Inspiration
IBM Moments of Time machine learning model, as well as the various uses possible if the model were more freely available
What it does
Uses a Convulutional Neural Network machine learning model to sift through an extremely large dataset to classify actions and implement itself into a variety of ways into bettering people’s lives as well as keeping them safer, all on a cohesive platform accessible via web application.
How I built it
Using the moments in times model, implemented a CNN based model to classify action and danger. Additionally, a node js based web application showcased several implementations and use cases of the model with a cohesive and modular interfaces
Challenges I ran into
Was finding an appropriate network for data manipulation, and working alone
Accomplishments that I'm proud of
The modularity and slickness of the final product made the model seamless and cohesive.
What I learned
Asynchronous operations in front end design and development, complex use of numerical data structures, and server based architectures
What's next for FistAction
Further develop the current models, and develop new models to further expand the reach of the tasks that the platform offers, one example being an educational/entertainment tool for the blind. Using the video action analysis machine model, I hope to develop accurate “subtitles for the blind”
Log in or sign up for Devpost to join the conversation.