Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·
 

Fast and accurate person re-identification with xception conv-net and C2F

Publication files not online:

Author: Rooijen, A.L. van · Bouma, H. · Verbeek, F.
Type:article
Date:2019
Publisher: Springer Verlag
Source:J.Morales, A. vera-rodriguez R.fierrez, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 23rd Iberoamerican Congress on Pattern Recognition, CIARP 2018, 19 November 2018 through 22 November 2018, 611-619
Identifier: 865906
ISBN: 9783030134686
Keywords: Convolutional neural networks · Feature extraction · Image retrieval · Large-scale person retrieval · Person re-identification · Cameras · Convolution · Feature extraction · Image retrieval · Neural networks · Coarse-to-fine searches · Convolutional neural network · Large-scale person retrieval · Lighting conditions · Multicamera systems · Person re identifications · Retrieval accuracy · State of the art · Pattern recognition

Abstract

Person re-identification (re-id) is the task of identifying a person of interest across disjoint camera views in a multi-camera system. This is a challenging problem due to the different poses, viewpoints and lighting conditions. Deeply learned systems have become prevalent in the person re-identification field as they are capable to deal with the these obstacles. Conv-Net using a coarse-to-fine search framework (Conv-Net+C2F) is such a deeply learned system, which has been developed with both a high-retrieval accuracy as a fast query time in mind. We propose three contributions to improve Conv-Net+C2F: (1) training with an improved optimizer, (2) constructing Conv-Net using a different Convolutional Neural Network (CNN) not yet used for person re-id and (3) coarse descriptors having fewer dimensions for improved speed as well as increased accuracy. With these adaptations Xception Conv-Net+C2F achieves state-of-the-art results on Market-1501 (single-query, 72.4% mAP) and the new, challenging data split of CUHK03 (detected, 42.6% mAP).