Skin lesion segmentation and classification using deep learning and handcrafted features

TL;DR

This paper introduces a hybrid feature approach combining handcrafted features with CNNs for skin lesion classification, significantly improving accuracy by 6.8% over traditional methods.

Contribution

It proposes a novel technique of injecting handcrafted features into CNNs during training, enhancing classification performance in dermoscopic image analysis.

Findings

Achieved 92.3% balanced multiclass accuracy

Hybrid features outperform single method features

Segmentation mask impacts classification accuracy

Abstract

Accurate diagnostics of a skin lesion is a critical task in classification dermoscopic images. In this research, we form a new type of image features, called hybrid features, which has stronger discrimination ability than single method features. This study involves a new technique where we inject the handcrafted features or feature transfer into the fully connected layer of Convolutional Neural Network (CNN) model during the training process. Based on our literature review until now, no study has examined or investigated the impact on classification performance by injecting the handcrafted features into the CNN model during the training process. In addition, we also investigated the impact of segmentation mask and its effect on the overall classification performance. Our model achieves an 92.3% balanced multiclass accuracy, which is 6.8% better than the typical single method classifier architecture for deep learning.