This paper presents an edge detection algorithm, using Haar wavelet transform with the maximum entropy thresholding technique on time-scale plane. The method was tested on 100 synthetic images at different signal-to-noise ratios (SNRs) and 7 clinical images. We assessed its reliability, accuracy and robustness using the MAD metric. The result for MAD on synthetic images yield MAD equal 0.7±0.14 and for clinical images give MAD equal 1.36±0.36 (mean±std).
Systems and Intelligent Techniques
