Melanoma Cancer Analysis Towards Early Detection Using Machine Learning Algorithms
In the recent years, computerized biomedical image and analysis are extremely promising and more interested and beneficial, it provides remarkable information and diagnoses of skin lesion. Melanoma distinguish one of lethal diseases. It is a type of skin cancer with highest suspicious
melanoma. The development modern diagnosis systems that can help in detecting of melanoma in its early stage to save life of many people. In this research work, the proposed system are highly promising. The novelty of this research work is to develop the effectiveness diagnosis system for
classification of melanoma and benign skin. The proposed system is divided into two phases, the first phase the classification algorithms have used to classify melanoma for identifying the melanoma. In the processing stage the Gaussian method was applied to improve the noise from dermosopy
images. The gaussian method was extremely important to help segmentation method for segmenting precise lesion from the skin images. The automatic segmentation has been considered to take out region of interest from dermosopy images for giving system more robust in detecting melanoma. As we
are working with PH2 data set that post-processing is required for improving the noise and hair that available in skin lesion, the morphological has used as post-processing. Subsequently, three features extraction methods have been implemented to extract lesion features, the Local Binary Pattern
(LBP) and Gray Level Co-occurrence Matrix (GLCM) methods have extract texture features, the statistical parameters have been used with GLCM methods where as the Fuzzy Color Histogram (FCH) method has extract color from lesion. These features have processed by classification algorithms for
diagnosis melanoma. Three classification algorithms namely Support Vector Machine (SVM), K-Nearest Neighbor (K-NN) and Artificial Neural Network (ANN) were applied, it has provided promising result. In second phase, ABCD (Asymmetry, Border, Color and Diameter) rules have been implemented for
automatic detecting melanoma. ABCD rules have four parameters to extract features from ROI image. After obtaining these features the Total Dermoscopic Score (TDS) is applied for detection melanoma. Obviously, we have used this method for validation the proposed system. It is observed that
the ABCD rules give proposed system more effect and efficiency for detecting the melanoma. Finally, it is proved that the proposed model has more efficiency and effectiveness for the diagnostic of skin cancer. Standard evaluation metrics like Accuracy, Specificity, Sensitivity, Precision,
Recall and F-score are employed to evaluate the results of proposed system. The evaluation and comparison the proposed system for classification and detection melanoma is presented. It is concluded that the proposed system is outperformed overall the existing systems.
Keywords: ABCD; CLASSIFICATION ALGORITHMS; FEATURE EXTRACTION; MELANOMA CANCER; SEGMENTATION
Document Type: Research Article
Publication date: 01 March 2019
- Journal of Medical Imaging and Health Informatics (JMIHI) is a medium to disseminate novel experimental and theoretical research results in the field of biomedicine, biology, clinical, rehabilitation engineering, medical image processing, bio-computing, D2H2, and other health related areas.
- Editorial Board
- Information for Authors
- Subscribe to this Title
- Ingenta Connect is not responsible for the content or availability of external websites
- Access Key
- Free content
- Partial Free content
- New content
- Open access content
- Partial Open access content
- Subscribed content
- Partial Subscribed content
- Free trial content