Image Enhancement for Image Matching Based Ant Colony Optimization using Update Pheromone Modification

To get the better result, Image matching needs better quality of method and image condition. But the problem is images are often corrupted with noise. The effect of this problem is quality of images decrease. A method to noise removal is proposed to enhance that image. The proposed method is combination between wavelet shrinkage and ant colony optimization (ACO) with modification on pheromone. The experiment uses six bitmap format images that grouped into three pairs of image. Denoising on three images shows difference result between image enhancement based ACO and image enhancement based ACO using update pheromone modification at 0.01 dB (PSNR). On other hand, denoising on another three images shows no difference PSNR value. For the amount of keypoint, the ground truths has 1756 keypoints, image on denoising based ACO has 1770 keypoints, and image on denoising based ACO using update pheromone modification has 1756. The resulting matching number shows the ground truth has average value at 95 correspondences, images on denoising using ACO has average value at 84 correspondences, and images on denoising using ACO using update pheromone modification has 79 correspondences.