{"slug":"sparsemap","name":"sparsemap","kind":"app","version":"0.1.1","description":"Live sparse feature map from your webcam: tracked corners, Lucas-Kanade optical flow, and a SLAM-style top-down camera trajectory + landmark map.","llm_description":"Turns a webcam into a live sparse map. One window overlays the feed with Good-Features-to-Track sparse feature points and Lucas-Kanade optical-flow vectors; a second window draws a SLAM-style top-down map of the camera trajectory plus the sparse landmarks those tracks accumulate (chained rigid 2D RANSAC estimation, planar assumption — a toy, not full 6-DoF SLAM). Single-file Python + OpenCV + numpy; run `python3 sparsemap.py` after `pip install -r requirements.txt`. Toggles for flow, points, stats, and map; keys r/f/p/m/s, space pauses, q quits.","keywords":["webcam","camera","sparse","map","optical-flow","feature-points","slam","trajectory","computer-vision","realtime"],"code_keywords":["python","opencv","numpy","pixi","conda-forge","lucas-kanade","good-features-to-track","estimateAffinePartial2D"],"languages":["python"],"author":"kenton","license":"MIT","homepage":"","entry_points":["sparsemap.py","README.md"],"install":"pixi install","run":"pixi run start","model":"GLM-5.3-Flash","prompt_summary":"User asked for a webcam project that renders a live sparsemap: sparse feature points, sparse optical flow, and a SLAM-style 2D trajectory/landmark map, in Python + OpenCV.","file_count":7,"total_bytes":367823,"downloads":18,"views":70,"created_at":"2026-10-17T05:00:19.954924486+00:00","updated_at":"2026-10-17T05:09:50.972706950+00:00","specs":null,"owner":"JG2HQZWButhcoh5/wf9vMtgFLIu0qocgLnJlqD6IWUM=","files":[".ndignore","README.md","notdouble.toml","pixi.lock","pixi.toml","requirements.txt","sparsemap.py"]}