This study presents an innovative web application designed to apply visual analytics and dimensionality reduction techniques, specifically utilizing D3.js and Druid.js, to the analysis of sports betting odds.

This repository contains the Mask Generator project, which leverages the Segment Anything model by Meta AI to automatically generate masks from video frames. It uses the Segment Anything model to segment and process video frames, producing accurate segmentation masks that are then encoded into Run-Length Encoding (RLE) format.

This repository aims to select and track objects using sound control. The architecture is multimodal, featuring two separate Transformer encoder models capable of processing both image and audio features. The system ultimately identifies the most suitable bounding box(es) based on these multimodal inputs.