Kandinsky LabKandinsky video analysis
Kandinsky Camera Motion Classifier
This multi-label video classifier adapts MCG-NJU's VideoMAE-Large. It predicts camera movement and shot-type labels for a whole input clip.
Research release
Use short segments for changing camera motion
The classifier samples sixteen frames across a clip. The card recommends two-second segments for longer videos because one whole-video sample can miss changes in movement.ai-forever/kandinsky-videomae-large-camera-motion — model card and files
Read the prediction at clip level
A label describes movement across the input segment. The model is not trained to return a label independently for every frame.ai-forever/kandinsky-videomae-large-camera-motion — model card and files
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | ai-forever/kandinsky-videomae-large-camera-motion |
| Base model | MCG-NJU/videomae-largeai-forever/kandinsky-videomae-large-camera-motion — model card and files |
| Labels | 18 motion classes and 3 shot-type classesai-forever/kandinsky-videomae-large-camera-motion — model card and files |
| Weight license | Not stated in the checked model card.ai-forever/kandinsky-videomae-large-camera-motion — model card and files |
Related models
- Kandinsky Video Motion Predictor
Kandinsky Lab · Kandinsky video analysis
A VideoMAEv2-Base adaptation that predicts three motion-intensity scores for video analysis.
Launch date not confirmedResearch releaseDownloadable