Image compression based on task-specific information

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

Compression is a key component in imaging systems hat have limited power, bandwidth, or other resources. In many applications, images are acquired to support a specific task such as target detection or classification. However, standard image compression techniques such as JPEG or JPEG2000 (J2K) are often designed to maximize image quality as measured by conventional quality metrics such as mean-squared error (MSE) or Peak Signal to Noise Ratio (PSNR). This mismatch between image quality metrics and ask performance motivates our investigation of image compression using a task-specific metric designed for the designated tasks. Given the selected target detection task, we first propose a metric based on conditional class entropy. The proposed metric is then incorporated into a J2K encoder to create compressed codestreams ha are fully compliant with the J2K standard. Experimental results illustrate that the decompressed images obtained using he proposed encoder greatly improve performance in detection/classification tasks over images encoded using a conventional J2K encoder.

Original languageEnglish (US)
Title of host publication2014 IEEE International Conference on Image Processing, ICIP 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages4817-4821
Number of pages5
ISBN (Electronic)9781479957514
DOIs
StatePublished - Jan 28 2014

Fingerprint

Image compression
Target tracking
Image quality
Imaging systems
Signal to noise ratio
Entropy
Bandwidth

Keywords

  • JPEG2000
  • Mutual Information
  • Target Detection
  • Task Specific Information

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Cite this

Pu, L., Marcellin, M. W., Bilgin, A., & Ashok, A. (2014). Image compression based on task-specific information. In 2014 IEEE International Conference on Image Processing, ICIP 2014 (pp. 4817-4821). [7025976] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICIP.2014.7025976

Image compression based on task-specific information. / Pu, Lingling; Marcellin, Michael W; Bilgin, Ali; Ashok, Amit.

2014 IEEE International Conference on Image Processing, ICIP 2014. Institute of Electrical and Electronics Engineers Inc., 2014. p. 4817-4821 7025976.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Pu, L, Marcellin, MW, Bilgin, A & Ashok, A 2014, Image compression based on task-specific information. in 2014 IEEE International Conference on Image Processing, ICIP 2014., 7025976, Institute of Electrical and Electronics Engineers Inc., pp. 4817-4821. https://doi.org/10.1109/ICIP.2014.7025976
Pu L, Marcellin MW, Bilgin A, Ashok A. Image compression based on task-specific information. In 2014 IEEE International Conference on Image Processing, ICIP 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 4817-4821. 7025976 https://doi.org/10.1109/ICIP.2014.7025976
Pu, Lingling ; Marcellin, Michael W ; Bilgin, Ali ; Ashok, Amit. / Image compression based on task-specific information. 2014 IEEE International Conference on Image Processing, ICIP 2014. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 4817-4821
@inproceedings{a0f306d855084d338023e919cfc92a44,
title = "Image compression based on task-specific information",
abstract = "Compression is a key component in imaging systems hat have limited power, bandwidth, or other resources. In many applications, images are acquired to support a specific task such as target detection or classification. However, standard image compression techniques such as JPEG or JPEG2000 (J2K) are often designed to maximize image quality as measured by conventional quality metrics such as mean-squared error (MSE) or Peak Signal to Noise Ratio (PSNR). This mismatch between image quality metrics and ask performance motivates our investigation of image compression using a task-specific metric designed for the designated tasks. Given the selected target detection task, we first propose a metric based on conditional class entropy. The proposed metric is then incorporated into a J2K encoder to create compressed codestreams ha are fully compliant with the J2K standard. Experimental results illustrate that the decompressed images obtained using he proposed encoder greatly improve performance in detection/classification tasks over images encoded using a conventional J2K encoder.",
keywords = "JPEG2000, Mutual Information, Target Detection, Task Specific Information",
author = "Lingling Pu and Marcellin, {Michael W} and Ali Bilgin and Amit Ashok",
year = "2014",
month = "1",
day = "28",
doi = "10.1109/ICIP.2014.7025976",
language = "English (US)",
pages = "4817--4821",
booktitle = "2014 IEEE International Conference on Image Processing, ICIP 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Image compression based on task-specific information

AU - Pu, Lingling

AU - Marcellin, Michael W

AU - Bilgin, Ali

AU - Ashok, Amit

PY - 2014/1/28

Y1 - 2014/1/28

N2 - Compression is a key component in imaging systems hat have limited power, bandwidth, or other resources. In many applications, images are acquired to support a specific task such as target detection or classification. However, standard image compression techniques such as JPEG or JPEG2000 (J2K) are often designed to maximize image quality as measured by conventional quality metrics such as mean-squared error (MSE) or Peak Signal to Noise Ratio (PSNR). This mismatch between image quality metrics and ask performance motivates our investigation of image compression using a task-specific metric designed for the designated tasks. Given the selected target detection task, we first propose a metric based on conditional class entropy. The proposed metric is then incorporated into a J2K encoder to create compressed codestreams ha are fully compliant with the J2K standard. Experimental results illustrate that the decompressed images obtained using he proposed encoder greatly improve performance in detection/classification tasks over images encoded using a conventional J2K encoder.

AB - Compression is a key component in imaging systems hat have limited power, bandwidth, or other resources. In many applications, images are acquired to support a specific task such as target detection or classification. However, standard image compression techniques such as JPEG or JPEG2000 (J2K) are often designed to maximize image quality as measured by conventional quality metrics such as mean-squared error (MSE) or Peak Signal to Noise Ratio (PSNR). This mismatch between image quality metrics and ask performance motivates our investigation of image compression using a task-specific metric designed for the designated tasks. Given the selected target detection task, we first propose a metric based on conditional class entropy. The proposed metric is then incorporated into a J2K encoder to create compressed codestreams ha are fully compliant with the J2K standard. Experimental results illustrate that the decompressed images obtained using he proposed encoder greatly improve performance in detection/classification tasks over images encoded using a conventional J2K encoder.

KW - JPEG2000

KW - Mutual Information

KW - Target Detection

KW - Task Specific Information

UR - http://www.scopus.com/inward/record.url?scp=84979073301&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84979073301&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2014.7025976

DO - 10.1109/ICIP.2014.7025976

M3 - Conference contribution

AN - SCOPUS:84979073301

SP - 4817

EP - 4821

BT - 2014 IEEE International Conference on Image Processing, ICIP 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -