COMPUTATION AND PERFORMANCE ANALYSIS OF DOUBLE STAGE FILTER FOR IMAGE PROCESSING

Autor:	Nurulfajar Abdul Manap, Masrullizam Mat Ibrahim, Teo Chee Huat
Rok vydání:	2015
Předmět:	Matching (graph theory) business.industry Computation ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION General Engineering Image processing Filter (signal processing) Dynamic programming Basic block Median filter Computer vision Segmentation Artificial intelligence business Mathematics
Zdroj:	Jurnal Teknologi. 77
ISSN:	2180-3722 0127-9696
DOI:	10.11113/jt.v77.6510
Popis:	Double Stage Filter (DSF) is a hybrid stereo matching algorithm which consists of basic block matching and dynamic programming algorithms, basic median filtering and new technique of segmentation. The algorithm acquire disparity maps which will be analyzed by using evaluation functions such as PSNR, MSE and SSIM. The computation of DSF and existing algorithms are presented in this paper. The Phase 2 in DSF is to remove the unwanted aspects such as depth discontinuities and holes from occlusion from the raw disparity map. Segmentation, merging and median filtering are the major parts for post processing of DSF algorithm. From the results of evaluation functions, the disparity maps attained by DSF is closer to the ground truth compared to other algorithms while its computation takes only few seconds longer than DP algorithm but its capable to obtain better results of disparity map.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::fe8a3ecd7c8fb210e9459a5f46a1664c https://doi.org/10.11113/jt.v77.6510 Zobrazit plný text záznamu