Pixlane is loading locally. No upload, no signup, and your files stay on your device.
Compute and visualize dense optical flow between two consecutive frames. See motion vectors, heatmaps showing displacement magnitude, and adjustable overlays.
Visualize dense optical flow between two frames — see motion vectors and displacement heatmaps.
All processing runs locally in your browser. Your files never leave your device — no upload, no server, no signup required.
Optical flow describes how each pixel appears to move between two frames captured a moment apart. This tool computes dense flow, producing a motion vector for every pixel rather than tracking only a handful of points. It uses the Farnebäck method, which approximates the neighborhood around each pixel with a local polynomial and then solves for the displacement that best explains how those polynomial coefficients change between the two frames. The result is a two-channel field: a horizontal and vertical displacement at every pixel. From that field the tool derives each pixel's magnitude (how far it moved) and angle (which direction), which is exactly what the heatmap and color-wheel visualizations encode.
Dense flow rests on the brightness-constancy assumption: a point keeps the same intensity as it shifts position. That holds for small, smooth motion under consistent lighting and breaks down under large jumps, occlusion, or sudden exposure changes. Because the estimate is built from local intensity gradients, it is closely related to edge detection — textured, high-contrast regions yield confident vectors, while flat areas like a clear sky or blank wall carry little gradient information and produce noisier flow.
Each output answers a different question, so pick the one that matches what you are inspecting:
For meaningful results the two frames must share the same viewpoint, with only the subject moving. If the whole image shifts because the camera moved, that global motion dominates the field; correcting it first with image alignment isolates the true subject motion. Unlike sparse approaches that follow individual keypoints from feature detection, dense flow characterizes the entire scene at once — what you want for motion heatmaps, segmentation cues, and stabilization analysis.
Dense optical flow underpins motion analysis: detecting movement in surveillance or wildlife footage, measuring fluid and crowd flow, stabilizing shaky clips, interpolating frames for slow motion, and separating a moving foreground from a static background. Students and researchers also use it to build intuition for how classical computer vision quantifies motion before reaching for heavier learned models.
Pixlane runs the entire computation in your browser with OpenCV compiled to WebAssembly. Your two frames never leave your device — nothing is uploaded to a server — so you can analyze private footage, medical imagery, or unreleased video with full confidence. It is free with no signup or watermark and returns results instantly, which makes it practical to sweep through many frame pairs and compare visualizations without per-image limits or network round trips.
Optical flow estimates the apparent motion of pixels between two consecutive images. It produces a dense vector field showing how each pixel has moved, which is fundamental in video analysis, motion tracking, and autonomous navigation.
Each motion vector shows the direction and magnitude of displacement for a pixel or region between the two frames. Longer arrows or brighter colors indicate larger movement.
The heatmap visualizes displacement magnitude — warmer colors (red, yellow) indicate areas with large motion, while cooler colors (blue, green) indicate little or no movement.
No. All processing runs locally in your browser using WebAssembly. Your images never leave your device — no upload, no server, no signup required.