PROJECT REPORT BACKGROUND AND FOREGROUND SEGMENTATION USING OTSU AND MANUAL THRESHOLDING
Daftar Isi:
- There are many image segmentation methods, one of them is thresholding. The problem is object and background segmentation on camouflage images is usually more difficult than discrete images. The solution in this project is result comparison of Otsu method which applied in two image type: discrete and camouflage. After histogram calculation of original image in each color channel, there is statistic calculation of Otsu method. The final step is searching for maximum value of between class variance, then it used as optimal threshold. While manual thresholding also provided to add flexibility on threshold values. The sliders are used as determination of threshold values from 0-255. Data sample which prepared for this research are 40 images that consists of 20 discrete images and 20 camouflage images. In general when tested using Otsu method or manual thresholding, the result is discrete images usually had better segmentation than camouflage images. Keywords: image segmentation, thresholding, otsu method, manual thresholding