Premium
A fuzzy and spline based dynamic histogram equalization for contrast enhancement of brain images
Author(s) -
Saravanan S.,
Karthigaivel R.
Publication year - 2021
Publication title -
international journal of imaging systems and technology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.359
H-Index - 47
eISSN - 1098-1098
pISSN - 0899-9457
DOI - 10.1002/ima.22483
Subject(s) - histogram equalization , adaptive histogram equalization , histogram , histogram matching , artificial intelligence , computer science , mathematics , balanced histogram thresholding , smoothing , color normalization , computer vision , pattern recognition (psychology) , image processing , image (mathematics) , color image
In general, medical images acquired at insufficient lighting conditions are suffering from low contrast issues which are inadequate for image analysis steps. One standard solution to improve the image contrast is to change its intensity distribution with the help of an image histogram. The histogram‐based contrast enhancement methods treat the images as regions rather than objects which would be more useful for applications like brain image enhancement. In this paper, a Fuzzy and Spline based Histogram Equalization (FSDHE) is proposed to perform contrast enhancement with medical images. The proposed FSDHE method partitions the image into connected components, and the type of components are identified with a fuzzy membership function. The dynamic histogram equalization is applied to each component individually. The equalized sub‐histograms are combined to drive the global histogram, which is inconsistent as dynamic histogram equalization treated the intensity range for each connected component differently. Hence, a spline‐based histogram smoothing is proposed here in this research work. The equalized intensity mapping is received as control points for the polynomial curve, and a smooth intensity transformation is interpolated as a spline curve. The proposed FSDHE model is analyzed with MRI‐brain image dataset of 3064 images, which consists of both benign and malignant cases. The contrast enhancement performance of the proposed FSDHE method is quantified by various measures like Absolute Mean Brightness Error (AMBE), Peak Signal to Noise Ratio (PSNR), Contrast (C), Weber Contrast (WC), Entropy, Hausdorff Distance (HD) and Texture Preservation (TP) measures. The performance of the FSDHE method is compared against with other histogram equalization methods, and the results indicate that the FSDHE method achieves better quality measures of 3.1401, 30.5499, 21.5486, 0.7779, 4.0252, 0.2777, and 0.7836 for the seven‐performance metrics.