Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET

Abstract: 

In video surveillance applications, person search is a chal-
lenging task consisting in detecting people and extracting
features from their silhouette for re-identification (re-ID) pur-
pose. We propose a new end-to-end model that jointly com-
putes detection and feature extraction steps through a single
deep Convolutional Neural Network architecture. Sharing
feature maps between the two tasks for jointly describing
people commonalities and specificities allows faster runtime,
which is valuable in real-world applications. In addition
to reaching state-of-the-art accuracy, this multi-task model
can be sequentially trained task-by-task, which results in a
broader acceptance of input dataset types. Indeed, we show
that aggregating more pedestrian detection datasets without
costly identity annotations makes the shared feature maps
more generic, and improves re-ID precision. Moreover, these
boosted shared feature maps result in re-ID features more
robust to a cross-dataset scenario.

up
0 users have voted:

Paper Details

Authors:
Angelique Loesch, Jaonary Rabarisoa, Romaric Audigier
Submitted On:
19 September 2019 - 12:16pm
Short Link:
Type:
Poster
Event:
Document Year:
2019
Cite

Document Files

2019_ICIP_aloesch

(4)

Keywords

Subscribe

[1] Angelique Loesch, Jaonary Rabarisoa, Romaric Audigier, "END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4743. Accessed: Oct. 18, 2019.
@article{4743-19,
url = {http://sigport.org/4743},
author = {Angelique Loesch; Jaonary Rabarisoa; Romaric Audigier },
publisher = {IEEE SigPort},
title = {END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET},
year = {2019} }
TY - EJOUR
T1 - END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET
AU - Angelique Loesch; Jaonary Rabarisoa; Romaric Audigier
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4743
ER -
Angelique Loesch, Jaonary Rabarisoa, Romaric Audigier. (2019). END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET. IEEE SigPort. http://sigport.org/4743
Angelique Loesch, Jaonary Rabarisoa, Romaric Audigier, 2019. END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET. Available at: http://sigport.org/4743.
Angelique Loesch, Jaonary Rabarisoa, Romaric Audigier. (2019). "END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET." Web.
1. Angelique Loesch, Jaonary Rabarisoa, Romaric Audigier. END-TO-END PERSON SEARCH SEQUENTIALLY TRAINED ON AGGREGATED DATASET [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4743