A DATASET AT A GLANCE...

Dataset for skin cancer diagnosis...

Skin cancers, the abnormal cell growth on the epidermis of the skin, has been increasing worldwide because of the incessant exposure to sunlight. Basal cell carcinoma, Melanoma, Squamous cell carcinoma and Merkel cell carcinoma are to name a few. Early diagnosis and treatment of skin cancer is utmost required, as the World Health Organization (WHO) and the International Labour Organization (ILO) have together estimated that about one out of three deaths caused by non-melanoma cancer is due to working under the sun. Further, as per the skin cancer foundation, an early diagnosis of melanoma can result in a 99% five-year survival rate. Though sunlight can be the major cause for skin cancer, other causes might include the Ultraviolet rays from tanning beds, immunosuppression and certain genetic disorders. Whatever might be the cause, if the skin cancer is left undiagnosed, it will start to metastasize and affect the inner organs, which is even more serious!!! Hence, research on skin cancer detection is much necessitated. In recent years, automatic detection of diseases with machine learning approaches have become more prevalent, as they provide precise detection results without expert interpretation. So, if a researcher is interested to lay approaches for automatic skin cancer diagnosis, he/she might use the ‘Skin Cancer ISIC’ dataset from Kaggle (https://www.kaggle.com/datasets/nodoubttome/skin-cancer9-classesisic). This dataset, encompassing about 2357 images of malignant and benign oncological diseases, is formed from the International Skin Imaging Collaboration (ISIC). With this dataset, about nine classes of skin cancer can be early- detected and they are: (i) actinic keratosis, (ii) basal cell carcinoma, (iii) dermatofibroma, (iv) melanoma, (v) nevus, (vi) pigmented benign keratosis, (vii) seborrheic keratosis, (viii) squamous cell carcinoma and (ix) vascular lesion. Moreover, this dataset is publicly available to use, without any citation request. 

Image courtesy: www.freepik.com