Advancing Skin Cancer Diagnosis: EfficientNetB1BasedDeep Learning Model for EnhanceAccuracy

Skin cancer is considered one of the most fatal kinds of cancer so, efficient early diagnosis techniques are essential to improve patient outcomes. Using the $\mathbf{1 0, 0 1 5}$ dermoscopic images from seven examination categories in the HAM10000 dataset -this work suggests a new deep learningbased method for the automated categorization of skin cancer. Using the EfficientNetB5 model, fine-tuned with extra layers and regularization methods, the system seeks to precisely identify benign from malignant skin lesions. Woman Among the several skin disorders the collection covers are Active keratoses, Basal cell carcinoma, Benign keratosis-like lesions, Dermatofibroma, Melanoma, Melanocytic Nevi, and Arterial Lesions. With an overall accuracy of 84% the suggested model was carefully trained and validated with recall, precision, and an F1 score of 0.86,0.84, and 0.85, respectively.Apart from the loss statistics, the model’s training and validation accuracy suggest robust learning and modest overfitting. These positive results show the viability of the EfficientNetB5-based approach to enhance early skin cancer detection, therefore providing clinicians with a reliable and rapid tool. Future studies will focus on overcoming the discovered limitations and looking at innovative technologies to raise the model’s even more diagnostic capacity.