1. Tutorial on Histogram Processing for Contrast Enhancement of Digital Images Brightness Preserving Contrast Enhancement Hrushikesh Garud Senior Software Engineer Texas Instruments (India), Bangalore and School of Medical Science and Technology, Indian Institute of Technology, Kharagpur International Conference on Data Engineering and Communication Systems (ICDECS-2011) 30-31 December 2011 Bangalore, India Thanks: Mr. Debdoot Sheet School of Medical Science and Technology, Indian Institute of Technology, Kharagpur
Fuzzy statistics of the digital images is in general used to effectively handle inexactness of the image data and to obtain a smooth histogram Smooth histogram helps perform its meaningful partitioning for brightness preserving equalization
where, high and low are the highest and lowest intensity value of the kth input sub-histogram and P_k is the total number of pixel in that partition
In eq.12: V’ is new gray value Start_k is starting intensity value for kth partition Range_k is range of kth partition as computed in previous sub-step h(i)/P_k is the probability of ith intensity value starting from start intensity value of that partition to intensity value v.
In the TABLE I: it can be seen that our technique (BPDFHE) Out-performs both HE and CLAHE and preserves the image brightness to the maximum extent
Here we compute the contrast of an image from rotational invariant Fuzzy-GLCM, obtained by averaging four symmetrical co-occurrence matrices obtained with different values of theta. In TABLE 2: it can be observed that BPDFHE provides contrast enhancement equivalent to HE and CLAHE, but it should be remembered that BPDFHE preserves the Manifestation of clinical feature and image brightness better than HE and CLAHE It has been observed that the BPDFHE provides contrast enhancement comparable to that provided by HE and CLAHE techniq ues. Whereas, it outperforms both He and CLAHE in image brightness preservation.
The non-shifting of the peaks in histogram helps to preserve the mean image-brightness while increasing contrast
HE though able to enhance the contrast but it leads to saturation of pixels to two extremities, even though overall contrast improves, the visibility of some of the details is lost CLAHE is able to enhance local contrast to large extent but it completely alters appearance of the different tissue regions in image, which may lead severe degradation in diagnostic value of the image. (such as regions of epithelial region) Where as BPDFHE enhances image contrast while preserving image brightness. It has been observed that while providing good contrast enhancement BPDFHE retains Manifestation of clinical feature (Texture of epithelial region /chromaticity (not to be confused with chroma inforation) of the nuclear regions.)
Another sample image, only lightness channel is considered Clearly notice the saturation effect in HE image. Whereas ClAHE and BPDFHE have comparable contrast enhancement