Evaluation of Intensity Transformation and Histogram-Based Methods for Image Enhancement

Abstract

Digital image enhancement is an important process in image processing because it improves visual quality and supports further image analysis. Images captured by digital devices often suffer from degradation caused by sensor limitations, uneven illumination, low contrast, and non-uniform intensity distribution. This study evaluates several grayscale image enhancement techniques based on intensity transformation and histogram analysis using an interactive Python-based application. The methods include image negative transformation, power-law transformation (gamma correction), gray level slicing, histogram equalization, histogram matching, and Contrast Limited Adaptive Histogram Equalization (CLAHE). Evaluation was performed through visual comparison, histogram analysis, and quantitative measurements, including mean intensity, entropy, Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), standard deviation, and Contrast Improvement Index (CII). The dataset consisted of six standard grayscale images from the Gonzalez and Woods image database: Lena, Cameraman, Boat, House, Peppers, and Baboon. All images were converted to grayscale and resized to 512 × 512 pixels.The results show that each method has different characteristics depending on the image condition. Histogram equalization effectively enhances global contrast but may cause over-enhancement. CLAHE provides more stable local contrast improvement and preserves image details. Power-law transformation offers flexible brightness adjustment, while gray level slicing highlights specific intensity ranges. The developed application supports real-time parameter adjustment and visualization of processed images, histograms, and evaluation metrics, providing an effective framework for image enhancement analysis and learning digital image processing