Standard: MPEG-AI
Part: 4

Editions

Edition - 1: Feature coding for machines

Publication Year: 0
Status: ongoing
Motivations: This project compresses the intermediate feature tensor(s) encountered within a machine vision task network itself, rather than input video to the machine vision task network, enabling applications where the computational burden of a neural network is to be performed in a distributed manner or higher coding efficiency compared to inferencing on decompressed video is required.
Objectives: Define bitstream syntax and semantics for compressed intermediate features that are efficient in terms of bitrate and enable connecting two parts of a machine vision task network without significantly reducing the quality of result of the machine vision task network.

Meeting documents

MPEG 153

Publication date	Title
0000-00-00	Common test and training conditions for FCM

MPEG 152

Publication date	Title
2025-11-22	Common test and training conditions for FCM

MPEG 151

Publication date	Title
2025-07-31	Common test and training conditions for FCM

MPEG 150

Publication date	Title
2025-05-04	Common test and training conditions for FCM

MPEG 149

Publication date	Title
2025-03-09	Common test and training conditions for FCM

MPEG 148

Publication date	Title
2024-12-31	Common test and training conditions for FCM

MPEG 147

Publication date	Title
2024-09-10	Common test and training conditions for FCM

MPEG 146

Publication date	Title
2024-06-12	Common test and training conditions for FCM

MPEG 145

Publication date	Title
2024-04-04	Common test and training conditions for FCM

MPEG 144

Publication date	Title
2024-01-02	Common test and training conditions for FCM