What
Add the ability to generate an animated GIF or MP4 that smoothly transitions from 0% to 100% intensity for a given procedure.
Why
Animated previews are much more compelling for patient consultations and for social media / demo purposes than static images. Seeing the gradual progression makes the prediction feel more trustworthy.
Suggested approach
- Generate frames at 2-5% intensity increments (20-50 frames)
- Assemble into GIF using
imageio or PIL
- Optionally output MP4 via
cv2.VideoWriter
- Add
--animate flag to CLI and a toggle in the Gradio demo
python scripts/run_inference.py photo.jpg \
--procedure rhinoplasty \
--animate \
--output rhinoplasty_progression.gif
Considerations
- Keep frame count reasonable for TPS mode (instant per frame)
- For ControlNet mode, this is expensive — consider caching the conditioning and only re-running the diffusion
- Add a progress bar for the generation loop
Difficulty: 🟡 Intermediate
What
Add the ability to generate an animated GIF or MP4 that smoothly transitions from 0% to 100% intensity for a given procedure.
Why
Animated previews are much more compelling for patient consultations and for social media / demo purposes than static images. Seeing the gradual progression makes the prediction feel more trustworthy.
Suggested approach
imageioorPILcv2.VideoWriter--animateflag to CLI and a toggle in the Gradio demopython scripts/run_inference.py photo.jpg \ --procedure rhinoplasty \ --animate \ --output rhinoplasty_progression.gifConsiderations
Difficulty: 🟡 Intermediate