Parameter-based transfer learning for severity classification of atopic dermatitis using hyperspectral imaging.

BACKGROUND/PURPOSE: Because atopic dermatitis (AD) is a chronic inflammatory skin condition that causes structural changes, there is a growing need for noninvasive research methods to evaluate this condition. Hyperspectral imaging (HSI) captures skin structure features by exploiting light wavelength variations in penetration depth. In this study, parameter-based transfer learning was deployed to classify the severity of AD using HSI. Therefore, we aimed to obtain an optimal combination of classification results from the four models after constructing different source- and target-domain datasets. METHODS: We designated psoriasis, skin cancer, eczema, and AD datasets as the source datasets, and the set of images acquired via hyperspectral camera as the target dataset for wavelength-specific AD classification. We compared the severity classification performances of 96 combinations of sources, models, and targets. RESULTS: The highest classification performance of 83% was achieved when ResNet50 was trained on the augmented psoriasis dataset as the source, with the resulting parameters used to train the model on the target Near-infrared radiation (NIR) dataset. The second highest classification accuracy of 81% was achieved when ResNet50 was trained on the unaugmented psoriasis dataset as the source, with the resulting parameters used to train the model on the target R dataset. ResNet50 demonstrated potential as a generalized model for both the source and target data, also confirming that the psoriasis dataset is an effective training resource. CONCLUSION: The present study not only demonstrates the feasibility of the severity classification of AD based on hyperspectral images, but also showcases combinations and research scalability for domain exploration.

Psoriasis severity classification based on adaptive multi-scale features for multi-severity disease.

Psoriasis is a skin disease that causes lesions of various sizes across the body and can persist for years with cyclic deterioration and improvement. During treatment, and a multiple-severity disease, with irregular severity within the observation area may be found. The current psoriasis evaluation is based on the subjective evaluation criteria of the clinician using the psoriasis area and severity index (PASI). We proposed a novel psoriasis evaluation method that detects representative regions as evaluation criteria, and extracts severity features to improve the evaluation performance of various types of psoriasis, including multiple-severity diseases. We generated multiple-severity disease images using CutMix and proposed a hierarchical multi-scale deformable attention module (MS-DAM) that can adaptively detect representative regions of irregular and complex patterns in multiple-severity disease analyses. EfficientNet B1 with MS-DAM exhibited the best classification performance with an F1-score of 0.93. Compared with the performance of the six existing self-attention methods, the proposed MS-DAM showed more than 5% higher accuracy than that of multiscale channel attention module (MS-CAM). Using the gradient-weighted activation mapping method, we confirmed that the proposed method works at par with human visual perception. We performed a more objective, effective, and accurate analysis of psoriasis severity using the proposed method.