4 ms·
Intuitively Understanding Harris Corner Detector
- arketyp 3y agoI remember the eigenvector analysis of the original paper [1] wasn't terribly inaccessible. I think an alternative title for this blog post could be "Intuitively Understanding the Harris Corner Detector Optimizations". [1] https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=88cdfbeb78058e0eb2613e79d1818c567f0920e2 https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&d...
- glitchc 3y agoI had to write one for an interview many moons ago. It's a fun little exercise.
- ur-whale 3y agoPSA : The Harris Corner detector, while interesting to understand if you like linear algebra, is not exactly what you'd call state of the art in the feature detection layer of computer vision.
- DougMerritt 3y agoYes? What is the state of the art?
- jhoydich 3y agoI believe the SIFT algorithm is most commonly used. Harris struggles when features change in scale between images, whereas SIFT does not. Harris can be outfitted with a Laplacian pyramid to overcome the scale issue though.
- mathisfun123 3y agowhich is funny since SIFT is basically just as ancient.
- anilz 3y agoActually, it is a part of the ORB algorithm and ORB is especially used a lot in visual SLAM applications.
- michaelt 3y agoThe most widely used algorithms for classical feature detection today are "whatever opencv implements" In terms of tech that's advancing at the moment? ML techniques. https://co-tracker.github.io/ https://co-tracker.github.io/ if you want to track individual points, https://github.com/matterport/Mask_RCNN https://github.com/matterport/Mask_RCNN and its descendents if you want to detect, say, the cover of a book.