Skin Lesion Extraction Using Multiscale Morphological Local Variance Reconstruction Based Watershed Transform and Fast Fuzzy C-Means Clustering

Early identification of melanocytic skin lesions increases the survival rate for skin cancer patients. Automated melanocytic skin lesion extraction from dermoscopic images using the computer vision approach is a challenging task as the lesions present in the image can be of different colors, there may be a variation of contrast near the lesion boundaries, lesions may have different sizes and shapes, etc. Therefore, lesion extraction from dermoscopic images is a fundamental step for automated melanoma identification. In this article, a watershed transform based on the fast fuzzy c-means (FCM) clustering algorithm is proposed for the extraction of melanocytic skin lesion from dermoscopic images. Initially, the proposed method removes the artifacts from the dermoscopic images and enhances the texture regions. Further, it is filtered using a Gaussian filter and a local variance filter to enhance the lesion boundary regions. Later, the watershed transform based on MMLVR (multiscale morphological local variance reconstruction) is introduced to acquire the superpixels of the image with accurate boundary regions. Finally, the fast FCM clustering technique is implemented in the superpixels of the image to attain the final lesion extraction result. The proposed method is tested in the three publicly available skin lesion image datasets, i.e., ISIC 2016, ISIC 2017 and ISIC 2018. Experimental evaluation shows that the proposed method achieves a good result.

Novel methods for automatic segmentation of abnormal lung parenchyma utilizing watershed and wavelet transforms

CT scan of the thorax is widely used to diagnose and evaluate numerous lung diseases. These scans yield a large amount of image data. The expanding volume of thoracic CT studies along with the increase of image data, elucidates the need of computer-aided diagnosis (CAD) schemes to assist the radiologists. Since several lung diseases are diagnosed based on the patterns of lung tissue in medical images, texture segmentation is an essential part of the [sic] most CAD systems. The processing step of most CAD systems is lung segmentation. In the first part of this thesis a novel approach for lung segmentation is proposed. The proposed method is based on watershed transform, which is fast and accurate. Lung region is precisely marked with internal and external markers. The markers are combined with the gradient image of the original data, then watershed transform is applied on the combined data to find the lung borders. A "Rolling ball" filter is used to fill the cavities and make the contour smooth while preserving the original borders. In the second part of this research work a novel composite method is proposed to segment the abnormality in lung tissue. The proposed approach is based on wavelet transform and intensity similarities. Our focus is on the honeycomb texture in lung tissue, which occurs with several interstitial lung diseases. After segmenting lung regions, Wavelet Transform is applied to decompose the image. The transformed lung region is thresholded to extract high resolution areas. Then the regions with low pixel intensities are kept and grown to segment the honeycomb regions. The proposed method has been tested on 91 pediatric chest CT images containing healthy and unhealthy lung images. Statistical analysis has been done and the results show the sensitivity of 100% along with the average Specificity of 95.14 %. A comparison with AMFM (82.5 % sensitivity and 99.9 % specificity) and ANN methods (100 % sensitivity and 88.1 % specificity) reveals the superiority of the proposed approach. The test results of both the lung segmentation and abnormal lung tissue segmentation techniques validate the robustness of the proposed methods.

Novel methods for automatic segmentation of abnormal lung parenchyma utilizing watershed and wavelet transforms

CT scan of the thorax is widely used to diagnose and evaluate numerous lung diseases. These scans yield a large amount of image data. The expanding volume of thoracic CT studies along with the increase of image data, elucidates the need of computer-aided diagnosis (CAD) schemes to assist the radiologists. Since several lung diseases are diagnosed based on the patterns of lung tissue in medical images, texture segmentation is an essential part of the [sic] most CAD systems. The processing step of most CAD systems is lung segmentation. In the first part of this thesis a novel approach for lung segmentation is proposed. The proposed method is based on watershed transform, which is fast and accurate. Lung region is precisely marked with internal and external markers. The markers are combined with the gradient image of the original data, then watershed transform is applied on the combined data to find the lung borders. A "Rolling ball" filter is used to fill the cavities and make the contour smooth while preserving the original borders. In the second part of this research work a novel composite method is proposed to segment the abnormality in lung tissue. The proposed approach is based on wavelet transform and intensity similarities. Our focus is on the honeycomb texture in lung tissue, which occurs with several interstitial lung diseases. After segmenting lung regions, Wavelet Transform is applied to decompose the image. The transformed lung region is thresholded to extract high resolution areas. Then the regions with low pixel intensities are kept and grown to segment the honeycomb regions. The proposed method has been tested on 91 pediatric chest CT images containing healthy and unhealthy lung images. Statistical analysis has been done and the results show the sensitivity of 100% along with the average Specificity of 95.14 %. A comparison with AMFM (82.5 % sensitivity and 99.9 % specificity) and ANN methods (100 % sensitivity and 88.1 % specificity) reveals the superiority of the proposed approach. The test results of both the lung segmentation and abnormal lung tissue segmentation techniques validate the robustness of the proposed methods.

Hematology and Digital Image Processing: Watershed Transform-Based Methodology for Blood Cell Counting Using the WT-MO Algorithm

Background: Most diseases can be detected by routine examination, even if they are in the initial phase. Currently, one of the most requested medical laboratory tests is that which allows detecting from bacterial infections until leukemias. However, for less favored populations, this examination can be seen as having a high cost. Methods: Thus, this study introduces an algorithm of segmentation of images capable of detecting and counting red blood cells and leukocytes present in digital images of blood smear. The methodology was named by WT-MO, once it relies on the concepts of Watershed Transform and Morphological Operations. The experiments were conducted in the MATLAB software simulation environment, where 25 images were used in order to evaluate the accuracy, processing time, and execution time of the WT-MO algorithm. Results: The results show that the WT-MO methodology presents high accuracy, reaching 96% and 92% in the red blood cell and leukocyte counts, respectively; reliability and low processing time, reaching an average processing time and execution time, achieving from 0.74 to 2.17 seconds. Therefore, the WT-MO algorithm can be seen as the first step in making laboratory tests more accessible to populations in underdeveloped and developing countries. Conclusion: The WT-MO methodology helps not only disadvantaged populations gain access to low-cost, high-reliability tests but also has excellent potential for use in laboratories in developed countries.

Superpixel-Based Shallow Convolutional Neural Network (SSCNN) for Scanned Topographic Map Segmentation

Motivated by applications in topographic map information extraction, our goal was to discover a practical method for scanned topographic map (STM) segmentation. We present an advanced guided watershed transform (AGWT) to generate superpixels on STM. AGWT utilizes the information from both linear and area elements to modify detected boundary maps and sequentially achieve superpixels based on the watershed transform. With achieving an average of 0.06 on under-segmentation error, 0.96 on boundary recall, and 0.95 on boundary precision, it has been proven to have strong ability in boundary adherence, with fewer over-segmentation issues. Based on AGWT, a benchmark for STM segmentation based on superpixels and a shallow convolutional neural network (SCNN), termed SSCNN, is proposed. There are several notable ideas behind the proposed approach. Superpixels are employed to overcome the false color and color aliasing problems that exist in STMs. The unification method of random selection facilitates sufficient training data with little manual labeling while keeping the potential color information of each geographic element. Moreover, with the small number of parameters, SCNN can accurately and efficiently classify those unified pixel sequences. The experiments show that SSCNN achieves an overall F1 score of 0.73 on our STM testing dataset. They also show the quality of the segmentation results and the short run time of this approach, which makes it applicable to full-size maps.