16 July 2013 Hidden Markov Models for background clutter
Author Affiliations +
The development of a target acquisition performance model for an electro-optical imaging system is seriously affected by the description of the target and background characteristics at present. Based on the Hidden Markov Model (HMM), a different clutter metric is proposed to quantify the influence of background on target detection in this article. It first simulates the process of recording a target in the human brain by optimizing the HMM parameters to represent the target as far as possible. And then the background clutter is defined to be the similarity, estimated by the computed model parameters, between the target and background. Finally, the newly proposed clutter metric is applied to the Search2 database, and the experiment results prove its superiority to other metrics.
© 2013 Society of Photo-Optical Instrumentation Engineers (SPIE)
Qian Li, Qian Li, Cui Yang, Cui Yang, Jian-Qi Zhang, Jian-Qi Zhang, Dong-Yang Zhang, Dong-Yang Zhang, } "Hidden Markov Models for background clutter," Optical Engineering 52(7), 073108 (16 July 2013). https://doi.org/10.1117/1.OE.52.7.073108 . Submission:

Back to Top