Logiciels

PIXANO

Intelligent opensource visual data annotation tool

LIMA

Linguistic analyzer (open source and proprietary), over 60 languages

Datasets

DAHLIA : DAily Human Life Activity

DAHLIA is a dataset for research into the recognition of activities of daily living.

MEL: Multimedia Entity Linking

Method for building a corpus for entity disambiguation, using visual and textual information

CCIHP : Characterized Crowd Instance-level Human Parsing

CCIHP is a dataset for research into the fine description of people using localized and characterized semantic attributes. It comprises 20 attribute classes and 20 feature classes, divided into 3 categories (size, pattern and color).

FALLMUD : FAscicle Lower Leg Muscle Ultrasound Dataset

FALLMUD consists of 812 ultrasound images of lower leg muscles annotated with muscle bundle and fascia masks.

H²O: Human-to-Human-or-Object Interaction Dataset

H²O is an image dataset for interaction detection. H²O is annotated according to a new interaction verb taxonomy including both person-object and person-person interactions.

Database for tracking rugby-7 players

Our dataset consists of three 40-second excerpts from matches at the World Rugby 2021 tournament in Dubai. Each video extract is annotated with player tracking data.

Open-set object detection: towards unified problem formulation and benchmarking (ECCV Workshop 2024)

We provide unified benchmarks (VOC-COCO and OpenImagesRoad) as well as codes for new evaluation metrics for Open-Set Object Detection according to the paper ‘Open-set object detection: towards unified problem formulation and benchmarking (ECCV Workshop 2024)’.