TitleFast and Scalable Keypoint Recognition and Image Retrieval Using Binary Codes
Publication TypeConference Paper
Year of Publication2011
AuthorsVentura, J., and T. Höllerer
Conference NameIEEE Workshop on Motion and Video Computing (WMCV 2011)

In this paper we report an evaluation of keypoint descrip- tor compression using as little as 16 bits to describe a single keypoint. We use spectral hashing to compress keypoint de- scriptors, and match them using the Hamming distance. By indexing the keypoints in a binary tree, we can quickly rec- ognize keypoints with a very small database, and efficiently insert new keypoints. Our tests using image datasets with perspective distortion show the method to enable fast key- point recognition and image retrieval with a small code size, and point towards potential applications for scalable visual SLAM on mobile phones.