Network latency remains the bottleneck for mobile visual search applications. We show how network latency can be reduced using Compressed Histogram of Gradient (CHoG) descriptors. We study the trade-off in Classification Accuracy (CA) and bitrate for different parameters of CHoG descriptors. We show how CHoG bitstreams can be used in a rate-scalable manner. The embedded representation of CHoG bitstreams reduces transmission delay and enables early termination on the server side. We obtain a 2-4 × decrease in system latency using CHoG descriptors compared to transmitting uncompressed SIFT descriptors or JPEG images in a 3G network.

Abstract Establishing visual correspondences is an essential component of many computer vision problems, which is often done with local feature-descriptors. Transmission and storage of these descriptors are of critical importance in the context of mobile visual search applications. We propose a framework for computing low bit-rate feature descriptors with a 20 × reduction in bit rate compared to state-of-theart descriptors. The framework offers low complexity and has significant speed-up in the matching stage. We show how to efficiently compute distances between descriptors in the compressed domain eliminating the need for decoding. We perform a comprehensive performance comparison with SIFT, SURF, BRIEF, MPEG-7 image signatures and other low bit-rate descriptors and show that our proposed CHoG descriptor outperforms existing schemes significantly over a wide range of bitrates. We implement the descriptor in a mobile image retrieval system and for a database of 1 million CD, DVD and book covers, we achieve 96 % retrieval accuracy using only 4 KB of data per query image.

In this paper we report an evaluation of keypoint descriptor compression using as little as 16 bits to describe a single keypoint. We use spectral hashing to compress keypoint descriptors, and match them using the Hamming distance. By indexing the keypoints in a binary tree, we can quickly recognize 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 keypoint recognition and image retrieval with a small code size, and point towards potential applications for scalable visual SLAM on mobile phones. 1.