International Journal of Engineering Research in Computer Science and Engineering (IJERCSE)

Skin Cancer Detection Using Texture Operators and Deep Learning Model

Author : K Preethika ¹ Kiran B S ² Kruthika Hebbar K K ³ Vinaya Narayan Moger ⁴ B. J. Sandesh ⁵ Prof. P Kokila ⁶

Date of Publication :31st July 2023

Abstract: Skin cancer cell detection can be done by extracting the relevant texture features. It can be more cost effective method when the dataset size is not relatively large and complex. Traditionally deep learning models have been used for classification of images as predictions have a high accuracy, however time taken to train a deep learning model is higher compared to a texture operator based model. Certain features of the images are relevant for accurate classification of images, when the dataset is less complex. Hence, a texture operator based model is developed which extracts relevant features using Gray Level Co-occurrence Matrix (GLCM). Colour features are also extracted and the features are classified using Light Gradient Boosting Machine (LightGBM). The result of the model is compared with the predictions of a deep learning model. The results are in good matching with deep learning model results.

Reference :

Will Updated soon