https://abaj.ai/tags/feature-matching/Recent content in Feature-Matching on Aayush Bajaj's Augmenting InfrastructureHugoen© 2026 Aayush BajajThu, 12 Nov 2026 02:04:46 +1100https://abaj.ai/wiki/computer-vision/sift/Wed, 11 Nov 2026 21:00:00 +1100https://abaj.ai/wiki/computer-vision/sift/<p>Take two photographs of the same building from slightly different spots. To glue them into one panorama you need three things: points that can be found again in both pictures, a way of deciding which point in one picture is which point in the other, and the single geometric map that carries one picture onto the other even though some of those decisions are wrong. The scale-invariant feature transform (SIFT) of Lowe (2004) supplies the first two; a homography fitted by random sample consensus (RANSAC) supplies the third.<span class="margin-note" data-note="Distils COMP9517 Week 3 (features, matching and alignment) and Lab 2, which the course ran on its own pair of photographs. Every number on this page was recomputed with OpenCV on the Lab 1 castle image (averaged over its ten noisy frames) and a second view made from it by a known homography."> <span class="margin-note-indicator">𐃏</span> </span> </p>