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

facebooktwittermailshare

Reconstruction-free deep convolutional neural networks for partially observed images

Abstract: 

Conventional image discrimination tasks are performed on fully observed images. In challenging real imaging scenarios, where sensing systems are energy demanding or need to operate with limited bandwidth and exposure-time budgets, or defective pixels, where the data collected often suffers from missing information, and this makes the task extremely hard. In this paper, we leverage Convolutional Neural Networks (CNNs) to extract information from partially observed images. While pre-trained CNNs fail significantly even with such a small percentage of the input missing, our proposed framework demonstrates the ability to overcome it after training on fully-observed and partially-observed images at a few observation ratios. We demonstrate that our method is indeed reconstruction-free, retraining-free and generalizable to previously untrained-on observation ratios and it remains effective in two different visual tasks – image classification and object detection. Our framework performs well even for test images with only 10% of pixels available and outperforms the reconstruct-then-classify pipeline in these challenging scenarios for small observation fractions.

up
0 users have voted:

Paper Details

Authors:
Arun Asokan Nair, Luoluo Liu, Akshay Rangamani, Peter Chin, Muyinatu A. Lediju Bell, Trac D. Tran
Submitted On:
26 November 2018 - 8:14pm
Short Link:
Type:
Presentation Slides
Event:
Document Year:
2018
Cite

Document Files

GlobalSIP_Poster_v2.pptx

(42)

Subscribe

[1] Arun Asokan Nair, Luoluo Liu, Akshay Rangamani, Peter Chin, Muyinatu A. Lediju Bell, Trac D. Tran, "Reconstruction-free deep convolutional neural networks for partially observed images", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3789. Accessed: Apr. 20, 2019.
@article{3789-18,
url = {http://sigport.org/3789},
author = {Arun Asokan Nair; Luoluo Liu; Akshay Rangamani; Peter Chin; Muyinatu A. Lediju Bell; Trac D. Tran },
publisher = {IEEE SigPort},
title = {Reconstruction-free deep convolutional neural networks for partially observed images},
year = {2018} }
TY - EJOUR
T1 - Reconstruction-free deep convolutional neural networks for partially observed images
AU - Arun Asokan Nair; Luoluo Liu; Akshay Rangamani; Peter Chin; Muyinatu A. Lediju Bell; Trac D. Tran
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3789
ER -
Arun Asokan Nair, Luoluo Liu, Akshay Rangamani, Peter Chin, Muyinatu A. Lediju Bell, Trac D. Tran. (2018). Reconstruction-free deep convolutional neural networks for partially observed images. IEEE SigPort. http://sigport.org/3789
Arun Asokan Nair, Luoluo Liu, Akshay Rangamani, Peter Chin, Muyinatu A. Lediju Bell, Trac D. Tran, 2018. Reconstruction-free deep convolutional neural networks for partially observed images. Available at: http://sigport.org/3789.
Arun Asokan Nair, Luoluo Liu, Akshay Rangamani, Peter Chin, Muyinatu A. Lediju Bell, Trac D. Tran. (2018). "Reconstruction-free deep convolutional neural networks for partially observed images." Web.
1. Arun Asokan Nair, Luoluo Liu, Akshay Rangamani, Peter Chin, Muyinatu A. Lediju Bell, Trac D. Tran. Reconstruction-free deep convolutional neural networks for partially observed images [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3789