Invention Grant
- Patent Title: Image compression selection based on interpretability loss estimation
-
Application No.: US15915160Application Date: 2018-03-08
-
Publication No.: US10827190B1Publication Date: 2020-11-03
- Inventor: Erik Philip Blasch , Hua-Mei Harry Chen , Zhonghai Wang , Genshe Chen , Kui Liu , Dan Shen
- Applicant: Government of the United States, as represented by the Secretary of the Air Force
- Applicant Address: US DC Washington
- Assignee: The United States of America as represented by the Secretary of the Air Force
- Current Assignee: The United States of America as represented by the Secretary of the Air Force
- Current Assignee Address: US DC Washington
- Agent Adam Gene Pugh
- Main IPC: H04N19/42
- IPC: H04N19/42 ; H04N19/85 ; G06T7/13

Abstract:
A framework for estimating image interpretability degradation associated with image compression is provided. An image compression broker system can determine an image compression setting to achieve an interpretability task in accordance with available communication bandwidth or transmission time objectives. Estimating image interpretability degradation includes detecting edge points of an uncompressed image and determining gradients corresponding to the detected edge points; compressing in accordance with a compression parameter setting the uncompressed image to generate a compressed image and determining gradients corresponding to the edge points in the compressed image; determining from the gradients associated with the edge points gradient ratios; and estimating from the gradient ratios an image interpretability loss of the compressed image.
Information query