Details

Autor(en) / Beteiligte

• Sonawane, Kavita
• Dsouza, Gavin
• Castillo, Oscar
• Medhane, Darshan V.
• Shinde, Swati V.

Titel

Vision Based Skin Cancer Detection: Various Approaches with a Comparative Study

Ist Teil von

• Applied Computer Vision and Soft Computing with Interpretable AI, 2024, Vol.1, p.35-53

Auflage

1

Ort / Verlag

United Kingdom: CRC Press

Erscheinungsjahr

2024

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Automatic diagnosis of skin cancer is one of the most challenging problems in medical image processing. It helps physicians to decide whether a skin melanoma is benign or malignant. So, determining the more efficient methods of detection to reduce the rate of errors is a vital issue among researchers. This automated diagnosis follows three important phases: preprocessing, feature extraction, followed by classification or detection. The proposed system focuses on the same flow with the application of thresholding-based segmentation to identify the region of interest and image enhancement as part of preprocessing. The color and texture contents are the main focus along with the geometry features in the second phase to strengthen the features. Machine learning is a promising field which is considered the state of the art for providing data insights. Thereby the application of emerging machine learning approaches for classification and detection fulfil the authors' purpose with the desired accuracy in detection. The experiment with the proposed system uses the dataset provided by the International Skin Imaging Collaboration Archive. The performance of the proposed model is evaluated using performance factors such as accuracy, precision, recall, and f1 score. The authors identified that logistic regression gives the best results with the combined feature vectors.