Video

Casual Conversations is a large scale multimodal (video + audio) benchmark dataset built to evaluate and audit computer vision and speech models for accuracy across diverse ages, genders, apparent skin tones, and lighting conditions.

Large-scale audio-visual dataset comprising speech clips with no interfering background signals.

These expressions are produced at two levels of emotional intensities (regular and strong) except for the neutral emotion that only contains regular intensity.

Each video contains some number of procedure steps to fulfill a recipe. All the procedure segments are temporal localized in the video with starting time and ending time. The distributions of 1) video duration, 2) number of recipe steps per video, 3) recipe segment duration and 4) number of words per sentence are shown below.

This corpus contains 15 spontaneous dialogues and multi-participant conversations by deaf signers, 10 of which were recorded in authentic settings like a deaf club and a bar, 5 were recorded in the lab.

Human lipreading performance increases for longer words, indicating the importance of features capturing temporal context in an ambiguous communication channel.

Data were collected from 100 subjects, yielding over thousand instances of synchronized data

The package including the videos and the metadata is available for non-commercial, academic research.

Real-World Activities of Daily Living

No one was asked to light a fire for this dataset. Video is taken from people's observations.

Licence plate recognition training model in Europe