brainstorming
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, modifying behavior, or planning anything new, in software or out of it (a talk, a business, a renovation).
video-and-trackingRun 3DDFA_V2 video, smoothing, and manual webcam tracking workflows.
Independent GitHub snapshot · Automated classification · Original documentation governs
Use this sub-skill for 3DDFA_V2 video-file processing and smoothing. Webcam support exists in the source repo but remains a manual, hardware-dependent workflow.
Read this sub-skill when the task asks to:
demo_video.py or demo_video_smooth.py on an AVI/MP4 file.2d_sparse, 2d_dense, or 3d video outputs.n_pre, n_next, start, or end frame-window settings.imageio, ffmpeg, tracking drift, or webcam/display issues.../setup-and-assets/.imageio and imageio-ffmpeg are installed.The basic wrapper preserves the original demo_video.py CLI. Put original
arguments after --:
python <skill-root>/sub-skills/video-and-tracking/scripts/run-video.py \
--repo-root <checkout> -- \
-f <video-path> -o 3d --onnx
The smoothed wrapper preserves the original demo_video_smooth.py CLI:
python <skill-root>/sub-skills/video-and-tracking/scripts/run-video-smooth.py \
--repo-root <checkout> -- \
-f <video-path> -o 2d_sparse -n_pre 1 -n_next 1 --onnx
n_pre + n_next + 1 frames and drops or
pads boundary frames.The repo has a webcam script that uses imageio.get_reader("<video0>") and
cv2.imshow. This skill documents the workflow in references/workflows.md,
but does not provide an automated webcam wrapper because it needs live camera
hardware, GUI display access, and manual quit handling.
Read references/troubleshooting.md for video-specific failure recovery and
../../references/troubleshooting.md for shared build/import failures.