Sorting Local Descriptors for Lowbit Rate Mobile Visual Search

Jie Chen,Ling-Yu Duan,Rongrong Ji,Hongxun Yao,Wen Gao

DOI: https://doi.org/10.1109/icassp.2011.5946582

2011-01-01

Abstract:State-of-the-art mobile visual search systems put emphasis on developing compact visual descriptors [4][6], which enables low bit rate wireless transmission instead of delivering an entire query image. In this paper, we address the orderless nature of the transmission set of query descriptors. We propose to adapt the orders of local descriptors in transmission, which subsequently yields more consistent statistic distributions in each feature dimension towards more efficient residual coding based compression. Our scheme further enables lossy sorting by an adaptive quantization strategy within each feature dimension, which largely improves the compression rates of the residual coding in each dimension. We show that the performance degeneration of such lossy sorting is acceptable in our mobile landmark search applications. Our approach's effectiveness and efficiency is demonstrated via extensive experimental comparisons to state-of-the art works in both mobile visual descriptors [2][4] and compact image signatures [5].