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

facebooktwittermailshare

End-to-end Learning Binary Representation via Direct Binary Embedding

Abstract: 

Learning binary representation is essential to large-scale computer vision tasks. Most existing algorithms require a separate quantization constraint to learn effective hashing functions. In this work, we present Direct Binary Embedding (DBE), a simple yet very effective algorithm to learn binary representation in an end-to-end fashion. By appending an ingeniously designed DBE layer to the deep convolutional neural network (DCNN), DBE learns binary code directly from the continuous DBE layer activation without quantization error. By employing the deep residual network (ResNet) as DCNN component, DBE captures rich semantics from images. Furthermore, in the effort of handling multilabel images, we design a joint cross entropy loss that includes both softmax cross entropy and weighted binary cross entropy in consideration of the correlation and independence of labels, respectively. Extensive experiments demonstrate the significant superiority of DBE over state-of-the-art methods on tasks of natural object recognition, image retrieval and image annotation.

up
0 users have voted:

Paper Details

Authors:
Hairong Qi
Submitted On:
16 September 2017 - 12:01pm
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Hairong Qi
Paper Code:
3028
Document Year:
2017
Cite

Document Files

Liu_icip_17.pdf

(211)

Subscribe

[1] Hairong Qi, "End-to-end Learning Binary Representation via Direct Binary Embedding", IEEE SigPort, 2017. [Online]. Available: http://sigport.org/2199. Accessed: Sep. 21, 2019.
@article{2199-17,
url = {http://sigport.org/2199},
author = {Hairong Qi },
publisher = {IEEE SigPort},
title = {End-to-end Learning Binary Representation via Direct Binary Embedding},
year = {2017} }
TY - EJOUR
T1 - End-to-end Learning Binary Representation via Direct Binary Embedding
AU - Hairong Qi
PY - 2017
PB - IEEE SigPort
UR - http://sigport.org/2199
ER -
Hairong Qi. (2017). End-to-end Learning Binary Representation via Direct Binary Embedding. IEEE SigPort. http://sigport.org/2199
Hairong Qi, 2017. End-to-end Learning Binary Representation via Direct Binary Embedding. Available at: http://sigport.org/2199.
Hairong Qi. (2017). "End-to-end Learning Binary Representation via Direct Binary Embedding." Web.
1. Hairong Qi. End-to-end Learning Binary Representation via Direct Binary Embedding [Internet]. IEEE SigPort; 2017. Available from : http://sigport.org/2199