Programs
- M. Tech. in Automotive Engineering -
- Clinical Fellowship in Laboratory Genetics & Genomics - Fellowship
Publication Type : Conference Proceedings
Publisher : IEEE
Source : 8th International Conference on Signal Processing and Integrated Networks (SPIN), Noida, India, 2021
Url : https://ieeexplore.ieee.org/abstract/document/9566039
Campus : Bengaluru
School : School of Artificial Intelligence
Verified : No
Year : 2021
Abstract : Psoriasis is an incurable and enduring skin disease. The standard metric used for the measurement of psoriasis severity and extent is known as Psoriasis Area and Severity Index (PASI). The dermatologists evaluate this score by using a sense of touch and visual examination, which is a manual, subjective, and tedious process. Therefore, in this paper, we proposed a deep learning framework for a fully automatic and objective process for measuring one of the most important parameters of PASI i.e. psoriasis area score. A modified U-Net model is implemented for the intended task. The proposed method automatically segments the lesion and healthy skin regions along with background simultaneously from the raw colour images of psoriasis patients from different body regions. This helps in the automatic measurement of psoriasis lesion area from the entire skin region in terms of percentage, which provides PASI area score. The study involves 350 images from 80 different psoriasis patients. The results of the proposed method are encouraging for the automatic estimation of the PASI area score. The proposed model has also achieved better performance compared to other state-of-the-art deep learning-based models.
Cite this Research Publication : Ritesh Raj, Narendra D. Londhe and Rajendra Sonawane, "Deep Learning based Multi-Segmentation for Automatic Estimation of Psoriasis Area Score," 2021 8th International Conference on Signal Processing and Integrated Networks (SPIN), Noida, India, 2021, pp. 1137-1142, https://doi.org/10.1109/SPIN52536.2021.9566039