Fast and accurate border detection in dermoscopy images using statistical region merging

Abstract

As a result of advances in skin imaging technology and the development of suitable image processing techniques during the last decade, there has been a significant increase of interest in the computer-aided diagnosis of melanoma. Automated border detection is one of the most important steps in this procedure, since the accuracy of the subsequent steps crucially depends on it. In this paper, a fast and unsupervised approach to border detection in dermoscopy images of pigmented skin lesions based on the Statistical Region Merging algorithm is presented. The method is tested on a set of 90 dermoscopy images. The border detection error is quantified by a metric in which a set of dermatologist-determined borders is used as the ground-truth. The proposed method is compared to six state-of-the-art automated methods (optimized histogram thresholding, orientation-sensitive fuzzy c-means, gradient vector flow snakes, dermatologist-like tumor extraction algorithm, meanshift clustering, and the modified JSEG method) and borders determined by a second dermatologist. The results demonstrate that the presented method achieves both fast and accurate border detection in dermoscopy images.

Texture Based Automated Segmentation of Skin Lesions using Echo State Neural Networks

“…Initialize weights in the reservoirs. [8] 89.4% 92.7% 92.3% Otsu-R [9] 87.3% 85.4% 84.9% Otsu-RGB [10] 93.6% 80.3% 80.2% Otsu-PCA [6] 79.6% 99.6% 98.1% TDLS [1] 91.2% 99.0% 98.3% Proposed Method 90.6% 99.2% 98.8%…”

“…Table 4 shows the detailed comparison of segmentation accuracy of the proposed method with five existing lesion segmentation techniques. The first algorithm (L-SRM) [8]. The second algorithm is (Otsu-RGB) [9], the third algorithm is (Otsu-PCA) the fourth algorithm is (TDLS) [1].…”

“…The majority of algorithms only use features derived from pixel color to drive the segmentation. Stoecker et al [18] proposed a novel texture based algorithm in skin images using basic statistical approaches such as the gray-level co-occurrence matrix and region merging [8]. From their algorithm, they found that texture analysis could accurately find regions with a smooth texture and it can be applicable for segmentation and classification of skin lesion images.…”

“…In this TD algorithm, the image is converted into XYZ color space. A TD metrics is calculated for each texture distribution based on the probability of it being similar to other texture distributions [8]. This information is combined with the contents of each region to determine a regional TD metric.…”

Multi-Organ Segmentation of CT Images using Statistical Region Merging

“…In [10] and [11] SRM was used for an automatic diagnosis of melanoma, in [12] for an automatic detection of breast cancer, and in [13] for evaluation of the post-operative outcome of knee prosthesis implantation.…”