A Novel Region-Based Modified Histogram Equalization for Enhancing Non-Uniformly Illuminated Chest X-Ray Images
Kirti Saini, Shivam Gangwar, Reeta Devi · Procedia Computer Science · 2025
A novel technique for improving non-uniformly illuminated chest X-ray (CXR’s) images is presented in this work. Current techniques based on histogram equalization (HE) might result in uneven illumination throughout the image and artificial looks and washed-out effects from over-enhancement. The suggested approach uses the exposure parameter to separate CXR images into three exposure regions: overexposed (OE) and underexposed (UE) and well exposed (WE) to solve this issue. A nonlinear weight in the histogram’s cumulative density function (or CDF) is used to alter each sub-region’s histogram. Then, utilizing altered transformation equations that offer distinct mapping directions and intensities expansion for underexposed, well exposed, and overexposed sub-regions, the updated histograms are equalized. Here we have used 1100 non-uniformly lighted CXR pictures from Kaggle dataset to assess the suggested method, and it was contrasted with three cutting-edge techniques: Brightness Preserving Bi-Histogram Equalization (BBHE) Dualistic Sub Image Histogram Equalization (DSIHE) and Histogram Equalization (HE). We have used here various parameters such as PSNR, ICF, AMBE to extract the more details of the proposed method with respect to three cutting-edge techniques. The suggested technique yielded pictures with more consistent lighting, improved detail preservation, and preserved naturalness. With the suggested approach, improvements were obtained that were both qualitative and quantitative.