Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic Images
Accurate classification of psoriasis is critical in dermatological diagnostics due to the disease’s diverse clinical presentations and varying severity levels. With numerous subtypes and their visual similarities to other dermatological conditions, precise diagnosis typically requires expert medica...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | Spanish |
| Published: |
Universidad Nacional de San Martín
2025-07-01
|
| Series: | Revista Científica de Sistemas e Informática |
| Subjects: | |
| Online Access: | https://revistas.unsm.edu.pe/index.php/rcsi/article/view/996 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849399987056672768 |
|---|---|
| author | Charu Bolia Sunil Joshi |
| author_facet | Charu Bolia Sunil Joshi |
| author_sort | Charu Bolia |
| collection | DOAJ |
| description |
Accurate classification of psoriasis is critical in dermatological diagnostics due to the disease’s diverse clinical presentations and varying severity levels. With numerous subtypes and their visual similarities to other dermatological conditions, precise diagnosis typically requires expert medical knowledge. Early and accurate identification of psoriasis subtypes is essential for initiating timely treatment. This study introduces a novel hybrid deep learning architecture that integrates EfficientNet with Long Short-Term Memory (LSTM) networks for the automated classification of psoriasis from dermoscopic images. The proposed model is designed to simultaneously capture spatial features through EfficientNet and temporal or sequential patterns via LSTM units, thereby improving classification performance. The models are trained and tested on a publicly benchmark dataset comprising 7 distinct classes using the publically available benchmark dataset by Dermnet and BFL-NTU. Experimental results demonstrate that the proposed architecture significantly outperforms the baseline models such as VGG16 and ResNet50, with superior accuracy 89.7% and robust performance across metrics like recall, F1-score with 88%, and Region of Convergence (ROC) of 97%. This compact design with low trainable parameters reduces the computational time and memory makes the model well-suited for deployment for portable devices and enabling real-time mobile-based dermatological assessments.
|
| format | Article |
| id | doaj-art-65fd1e0279d84d5fb5992deca7901682 |
| institution | Kabale University |
| issn | 2709-992X |
| language | Spanish |
| publishDate | 2025-07-01 |
| publisher | Universidad Nacional de San Martín |
| record_format | Article |
| series | Revista Científica de Sistemas e Informática |
| spelling | doaj-art-65fd1e0279d84d5fb5992deca79016822025-08-20T03:38:12ZspaUniversidad Nacional de San MartínRevista Científica de Sistemas e Informática2709-992X2025-07-015210.51252/rcsi.v5i2.996Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic ImagesCharu Bolia0https://orcid.org/0009-0009-2042-471XSunil Joshi 1Maharana Pratap University of Agriculture and Technology Maharana Pratap University of Agriculture and Technology Accurate classification of psoriasis is critical in dermatological diagnostics due to the disease’s diverse clinical presentations and varying severity levels. With numerous subtypes and their visual similarities to other dermatological conditions, precise diagnosis typically requires expert medical knowledge. Early and accurate identification of psoriasis subtypes is essential for initiating timely treatment. This study introduces a novel hybrid deep learning architecture that integrates EfficientNet with Long Short-Term Memory (LSTM) networks for the automated classification of psoriasis from dermoscopic images. The proposed model is designed to simultaneously capture spatial features through EfficientNet and temporal or sequential patterns via LSTM units, thereby improving classification performance. The models are trained and tested on a publicly benchmark dataset comprising 7 distinct classes using the publically available benchmark dataset by Dermnet and BFL-NTU. Experimental results demonstrate that the proposed architecture significantly outperforms the baseline models such as VGG16 and ResNet50, with superior accuracy 89.7% and robust performance across metrics like recall, F1-score with 88%, and Region of Convergence (ROC) of 97%. This compact design with low trainable parameters reduces the computational time and memory makes the model well-suited for deployment for portable devices and enabling real-time mobile-based dermatological assessments. https://revistas.unsm.edu.pe/index.php/rcsi/article/view/996skin diseaselong short term memorypsoriasisefficient net |
| spellingShingle | Charu Bolia Sunil Joshi Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic Images Revista Científica de Sistemas e Informática skin disease long short term memory psoriasis efficient net |
| title | Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic Images |
| title_full | Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic Images |
| title_fullStr | Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic Images |
| title_full_unstemmed | Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic Images |
| title_short | Optimized Deep Neural Network for High-Precision Psoriasis Classification from Dermoscopic Images |
| title_sort | optimized deep neural network for high precision psoriasis classification from dermoscopic images |
| topic | skin disease long short term memory psoriasis efficient net |
| url | https://revistas.unsm.edu.pe/index.php/rcsi/article/view/996 |
| work_keys_str_mv | AT charubolia optimizeddeepneuralnetworkforhighprecisionpsoriasisclassificationfromdermoscopicimages AT suniljoshi optimizeddeepneuralnetworkforhighprecisionpsoriasisclassificationfromdermoscopicimages |