Simple visual tracker + annotation UI for microscopy / video sources

HoloTrack is a small Python/OpenCV project that combines an easy-to-use Tkinter frontend with a robust, practical tracking pipeline. It’s designed for annotating regions in a video (e.g., microscopy) and keeping those annotations roughly locked to the object as it moves using a mixture of template-matching, Lucas–Kanade optical flow, and periodic feature replenishment.

This README explains what the project does, how to run it, and the most important knobs you may want to tweak.

• Launch a compact GUI to pick a video file (or use the default Microscopy1.mp4).
• Drag the mouse to draw contour regions you want to track over time.
• The tracker uses local features + optical flow to estimate object motion and warps your drawn contours accordingly.
• If the tracker loses the object it attempts to re-acquire it using template matching inside a search window.

• Lightweight Tkinter GUI for quick experiments.
• Click-to-anchor & draw-to-annotate workflow.
• Robustness measures:
  • Replenishes feature points when they get sparse.
  • Detects when the tracker is lost and tries to re-find the object with template matching.
  • Applies an estimated affine transform to keep drawn contours aligned with motion.

• Python 3.8+
• opencv-python
• numpy

Install dependencies with pip:

pip install opencv-python numpy

Click here to access the demo