Hysteresis thresholding based edge detectors for inscriptional image enhancement

Objective of the study is to suggest a hybrid computational method to enhance text from historical inscriptional images by differentiating foreground and background of the inscriptional images using hysteresis thresholding based on various edge detectors such as canny, sobel and laplacian techniques. Furthermore, the combined thresholding methods based on edge detectors were estimated to determine the peak signal to noise ratio (PSNR) and mean absolute difference (MAD) for the performance measurement error. Based on the enhancement results of 75 images, the combined hysteresis with global threshold based canny edge detector yields a better enhancement of text over individual threshold based edge detectors.