Browsing by Author "Yang, Yi-Hong"

Now showing 1 - 3 of 3

Simple median-based EP PP scheme for enhancement of reconstructed Bayer colour filter array images
(2016-12)
Yang, Yi-Hong
;
Wang, Peng-Hua
;
Chen, Po-Ning
;
電機工程學系
;
Department of Electrical and Computer Engineering
In this study, the median-based edge-preserving (EP) modification is revisited and has been shown to improve effectively the quality of a reconstructed Bayer colour filter array image in terms of the composite peak signal-to-noise ratio (CPSNR) performance index. Along this research direction, the authors propose an EP-modified signal-correlation-based (SCB) post-processing (PP), called EP-SCB, as an enhancement to any existing interpolation method. The Bayer image reconstruction system they consider thus consists of two operational phases. The first phase performs an initial estimation of the missing red, green and blue colours by using an existing interpolation method, whereas the second phase applies EP-SCB PP. Since a certain class of images may not fulfil the premise of having small variations in local colour difference, which is assumed by SCBtype interpolation, thereby resulting in a deterioration in CPSNR after EP-SCB PP, a threshold test on local variance ratio is also devised to conditionally switch off the second phase. Experimental results show that the EP-SCB PP with variance-ratio test gives a worst average CPSNR than the original interpolation methods tested in none of the Kodak and IMAX image groups experimented.
Simple median-based EP PP scheme for enhancement of reconstructed Bayer colour filter array images
(2016-12)
Yang, Yi-Hong
;
Wang, Peng-Hua
;
Chen, Po-Ning
;
電機工程學系
;
Department of Electrical and Computer Engineering
In this study, the median-based edge-preserving (EP) modification is revisited and has been shown to improve effectively the quality of a reconstructed Bayer colour filter array image in terms of the composite peak signal-to-noise ratio (CPSNR) performance index. Along this research direction, the authors propose an EP-modified signal-correlation-based (SCB) post-processing (PP), called EP-SCB, as an enhancement to any existing interpolation method. The Bayer image reconstruction system they consider thus consists of two operational phases. The first phase performs an initial estimation of the missing red, green and blue colours by using an existing interpolation method, whereas the second phase applies EP-SCB PP. Since a certain class of images may not fulfil the premise of having small variations in local colour difference, which is assumed by SCBtype interpolation, thereby resulting in a deterioration in CPSNR after EP-SCB PP, a threshold test on local variance ratio is also devised to conditionally switch off the second phase. Experimental results show that the EP-SCB PP with variance-ratio test gives a worst average CPSNR than the original interpolation methods tested in none of the Kodak and IMAX image groups experimented.
Sufficient condition on hypothesis statistics for suboptimality of the identical-quantizer parallel distributed detection system
(2018-01-01)
Yang, Yi-Hong
;
Wang, Peng-Hua
;
Chen, Po-Ning
;
電信工程研究所
;
Institute of Communications Engineering
A challenge in designing an optimal parallel distributed detection system is its exponentially increasing number of possible combinations of local quantization rules with respect to n local sensors. Instead, one may focus on a much easier task of finding the best identical-quantizer system (IQS). This raises an important query about when the simple best IQS is globally optimal in Bayes detection error, particularly under a finite number of sensors. We consider this issue by asking the different but relevant question of when the best IQS is only suboptimal. By numerical experiments over the simplest system setting with ternary local observations and binary local quantizers, we find and prove that for every even n >= 2, if the output statistics of a local quantizer are symmetric with respective to the two hypotheses, the corresponding IQS design can never be optimal in the sense of minimizing the Bayes detection error subject to equal prior probabilities. Our analysis also indicates that asymptotically, one can almost halve the detection error of an IQS by replacing one of the identical local quantizers with a different one when the two hypothesis distributions place most of their probability masses, respectively, on two different local observation outcomes.