Project - European Data Incubator -- Unlocking data-driven innovation for European SMEs
Removing barriers to data access and innovation to help SMEs grow within Europe’s data economy
- Financed by
- Horizon 2020
- Duration
- Keywords
- Media AI Indexing & Discovery
Overview
IN2 has been selected for the European Data Incubator (EDI) programme, as one of the brightest Big Data startups in Europe. EDI gives the most innovative companies the opportunity to tackle real-world challenges set by corporates across Europe.
After pitching to a jury of technical experts, business experts and investors, we were chosen as one of the top 18 startups and SMEs from a pool of 208. We are going to solve the challenge set by VRT: Automatic Chaptering of Video Content.
Our approach is to use AI algorithms in order to extract features from the different modalities of the audiovisual material and use big data for learning a temporal structural model to represent a specific broadcasting format, starting with news bulletins. In this way, IN2’s proposed solution will automatically and accurately extract topics using features from both image, video and audio information and thus make it possible for broadcasters that produce large amounts of content to make it easily retrievable both by their audience and editing teams.
The project builds upon the MyMeedia Platform, IN2’s intelligent, modular, flexible content platform for orchestrating processing pipelines, including for:
- Text processing: language detection, stop-word elimination, tag-cloud computation, document classification, topic detection, named entity detection
- Image processing: color descriptors, texture descriptors, object detection, classification, object recognition
- Video processing: scene segmentation, camera motion, person detection, generic concept training and detector using machine learning approaches
- Audio processing: audio segmentation, classification (speech/non-speech, background), speaker classification and automatic speech recognition
The main outcome of IN2's participation in EDI is Vide, a Customised Fast Video Processing Engine.
Our role
- Feature Extraction from audio, image and video information
- Multimodal AI-content processing pipelines
- Development of temporal structual models for brodcasting news
- Topic detection and keyword extraction
- Pilots and Applications with the project's end users
