Final Year Group Project : Video
Text-to-Video
Audio Conditioned LipSync with Latent Diffusion Models
Extract audio, transcribe, and chunk YouTube video
Generate lifelike video animations from images and audio
Video Super-Resolution with Text-to-Video Model
Generate realistic talking heads from image+audio
Robotics Language-Gesture Video Generation
Generate animated characters from images
Stream audio/video in realtime with webrtc
a super consistent video depth model
Creator Friendly Text-to-Video
Generate responses to video or image inputs
Team18 Video Project is a Final Year Group Project focused on video generation and analysis. It is designed to identify driver distractions from video footage, making it a valuable tool for automotive safety and research. The project leverages AI technology to analyze videos and detect potential distractions, providing insights into driver behavior.
What types of distractions can Team18 Video Project detect?
Team18 Video Project can detect a variety of driver distractions, including texting, eating, talking to passengers, and reaching for objects.
Can Team18 Video Project work with real-time video feeds?
Yes, the project supports real-time video analysis, making it suitable for live monitoring applications.
How accurate is the distraction detection system?
The accuracy of the system depends on the quality of the video and lighting conditions but is optimized for high precision in typical driving scenarios.