Aster describes what a lecture video shows but never says, so blind and low-vision students can follow it
Detecção de cenas para que deficientes visuais consigam saber tudo que tem ao seu redor e suas respectivas posições no ambiente.
A system for generating music playlists based on the results of audio content analysis. The MusAV dataset is used as a music audio collection, music descriptors are extracted using Essentia, and a simple user interface is used to generate playlists based on these descriptors.
Omni Describer — AI-powered audio description for video, images and PDFs. Accessible descriptions for blind and visually impaired users.
Script and manifest files to create an HLS program containing the Elephants Dream video with captions, subtitles, and audio description
One voice at a time. An audio director for Fire TV that lets audio description, spoken subtitles and Baton's own spoken menus take turns, so a blind viewer never hears two voices at once. Kotlin, Media3, Compose for TV.
Sync fan-made audio description tracks with your video files. Accessibility-first, one command.
Find, play and save audio and video from about 1,750 sites, on Mac and Windows. Works the same by ear and by eye.
"DANTE-AD: Dual-Vision Attention Network for Long-Term Audio Description" CVPR Workshop AI4CC 2025
DescribeAT Admin Portal - content management for audio descriptions (React/Vite). Public open-source release.
Automated Audio Description (AD) sync toolkit with acoustic cross-correlation, DTW alignment, ITU-R BS.775 2.0 downmixing, and commercial excision.
Code release for NeurIPS 2026 paper "Mind The Gap (MTG-AD): Dataset and Fine-grained Evaluation for End-to-End Inline Audio Descriptions".
One video, two audio tracks, two output devices - so people who need different languages, or audio description, can watch the same screen together.
One video, made for every viewer: audio description (with extended description), captions with sounds, easy-read, and Hindi, Italian, Spanish, French and Portuguese versions. Gemma 4 31B on AMD Token Factory writes, code checks every line, a person approves.
DescribeAT - accessible audio description app (React/Vite). Public open-source release.
AI audio description for Fire TV (Vega OS): Amazon Polly voices descriptions generated from the film's frames, spoken in the gaps between dialogue.
Fire TV app that softens flashes, caps sudden loud sounds, warns before scares and describes scenes on your phone, for everyone on the couch.
Audio description for Fire TV, generated for films that do not have it — and never spoken over the dialogue
Audio-first football match companion for blind and low-vision fans: verified audio description from synthetic match events. Built for the Microsoft x Premier League Inside the Game hackathon.
To associate your repository with the audio-description topic, visit your repo's landing page and select "manage topics."