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

facebooktwittermailshare

AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS

Abstract: 

We address the task of estimating depth from a single intensity image via a novel convolutional neural network (CNN) encoder-decoder architecture, which learns the depth information using example pairs of color images and their corresponding depth maps. The proposed model integrates residual connections within pooling and up-sampling layers, and hourglass networks which operate on the encoded features, thus processing these at various scales. Furthermore, the model is optimized under the constraints of perceptual as well as the mean squared error loss. The perceptual loss considers the high-level features, thus operating at a different scale of abstraction, which is complementary to the mean squared error
loss. The improvements in qualitative and quantitative comparisons with state-of-the-art approaches demonstrate the effectiveness of our approach, even in the presence of noise.

up
0 users have voted:

Paper Details

Authors:
Seema Kumari, Ranjeet Ranjhan Jha, Arnav Bhavsar and Aditya Nigam
Submitted On:
19 September 2019 - 12:58am
Short Link:
Type:
Poster
Event:
Presenter's Name:
Ankit Dave
Paper Code:
MA.PB.8
Document Year:
2019
Cite

Document Files

Poster

(25)

Keywords

Additional Categories

Subscribe

[1] Seema Kumari, Ranjeet Ranjhan Jha, Arnav Bhavsar and Aditya Nigam, "AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4653. Accessed: Dec. 08, 2019.
@article{4653-19,
url = {http://sigport.org/4653},
author = {Seema Kumari; Ranjeet Ranjhan Jha; Arnav Bhavsar and Aditya Nigam },
publisher = {IEEE SigPort},
title = {AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS},
year = {2019} }
TY - EJOUR
T1 - AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS
AU - Seema Kumari; Ranjeet Ranjhan Jha; Arnav Bhavsar and Aditya Nigam
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4653
ER -
Seema Kumari, Ranjeet Ranjhan Jha, Arnav Bhavsar and Aditya Nigam. (2019). AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS. IEEE SigPort. http://sigport.org/4653
Seema Kumari, Ranjeet Ranjhan Jha, Arnav Bhavsar and Aditya Nigam, 2019. AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS. Available at: http://sigport.org/4653.
Seema Kumari, Ranjeet Ranjhan Jha, Arnav Bhavsar and Aditya Nigam. (2019). "AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS." Web.
1. Seema Kumari, Ranjeet Ranjhan Jha, Arnav Bhavsar and Aditya Nigam. AUTODEPTH: SINGLE IMAGE DEPTH MAP ESTIMATION VIA RESIDUAL CNN ENCODER-DECODER AND STACKED HOURGLASS [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4653