Discriminative Quantization for Fast Similarity Search

Sepehr Eghbali, Ladan Tahvildari; The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2019, pp. 0-0

Abstract


Recent decade has witnessed a growing surge of research on encoding high-dimensional objects with compact discrete codes. In this paper, we present a new supervised quantization technique to learn discriminative and compact codes for large scale retrieval tasks. To achieve fast and accurate search, the proposed algorithm learns a discriminative embedding of the input points and at the same time encodes the embedded points with compact codes to reduce storage cost.

Related Material


[pdf]
[bibtex]
@InProceedings{Eghbali_2019_CVPR_Workshops,
author = {Eghbali, Sepehr and Tahvildari, Ladan},
title = {Discriminative Quantization for Fast Similarity Search},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {June},
year = {2019}
}