Skin-Cancer-Detection-Using-Hybrid-Model

I have explored the efficacy of deep neural networks in automating skin cancer identification, addressing challenges like classification complexity and dataset scarcity. Using two publicly available datasets consisting 3309 and 5000 photos, I combined established deep learning models such as ResNet-50, VGG16, ShuffleNet, MobileNet, and DenseNet-201, along with novel hybrid structures. Data preprocessing includes steps like splitting, normalization, reshaping, and encoding, followed by augmentation to increase dataset volume. Traditional deep learning models and machine learning classifiers are independently evaluated, then re-evaluated in combination. Hybrid models, paired with Support Vector Machine (SVM) classifiers, outperform traditional DL models and ML classifiers. Notably, the hybrid VGG16 and ResNet50 with SVM classifier achieve highest accuracies on the respective datasets, indicating promise for early skin cancer detection.
DATASET 1
https://www.kaggle.com/datasets/fanconic/skin-cancer-malignant-vs-benign
DATASET 2
https://www.kaggle.com/datasets/hasnainjaved/melanoma-skin-cancer-dataset-of-10000-images/data