Efficient Adaptive Segmentation Based Contrast Enhancement and Image Binarization

• Main Authors: Chia-Shao Hung, 洪嘉劭
• Other Authors: Shanq-Jang Ruan
• Format: Others
• Language: en_US
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/45569033422887708435

博士 === 國立臺灣科技大學 === 電子工程系 === 102 === Image segmentation has been used widely in the area of image processing nowadays.Aiming at representing input image in a way which is easier to analysis, the precess makes the collection of local information for accurate calculation possible.However, when using this kind of technique on contrast enhancement or image binarization, traditional segmentation processes will cause the decrement of result quality due to ignore nearby pixels on edge pixels.In order to solve the problem, we propose intelligent block segmentation (IBS) based methods on both contrast enhancement and image binarization. Contrast enhancement involves transforming the intensity from the original state to feature significant impact on display devices.When using IBS on this kind of algorithms, the proposed one can provide both high quality and low edge loss rate.Firstly, the image will be segmented into different sized sub-images according to pixel characteristics.The local transform function of each sub-image will then be calculated by Bilateral Bezier Curve histogram equalization method.Once all local functions are calculated, the global transform function can be estimated by combining all functions together with weighting. Image binarization, however, is a process of converting gray-level into binary ones.The related algorithms can be classified as either high quality computation or high performance.After the input image is segmented, each sub-image will be classified as belonging to foreground or background ones.To sum up, by using IBS the result quality of both methods are increased.Experimental results also reveal that these methods can both providing high quality and effectiveness.