Data-Driven Deep Supervision for Skin Lesion Classification
This work addresses classification accuracy issues in medical imaging for skin lesions, representing an incremental improvement in deep learning methods for this domain.
The paper tackles the problem of robust feature extraction for skin lesion classification by proposing a deep neural network that uses activation mapping to select the optimal layer for deep supervision based on the match between the layer's receptive field and the approximated object shape, achieving verified effectiveness on five datasets.
Automatic classification of pigmented, non-pigmented, and depigmented non-melanocytic skin lesions have garnered lots of attention in recent years. However, imaging variations in skin texture, lesion shape, depigmentation contrast, lighting condition, etc. hinder robust feature extraction, affecting classification accuracy. In this paper, we propose a new deep neural network that exploits input data for robust feature extraction. Specifically, we analyze the convolutional network's behavior (field-of-view) to find the location of deep supervision for improved feature extraction. To achieve this, first, we perform activation mapping to generate an object mask, highlighting the input regions most critical for classification output generation. Then the network layer whose layer-wise effective receptive field matches the approximated object shape in the object mask is selected as our focus for deep supervision. Utilizing different types of convolutional feature extractors and classifiers on three melanoma detection datasets and two vitiligo detection datasets, we verify the effectiveness of our new method.