Application of ANN in discriminating skin lesions / Roziah Jarmin
This work describes the development of a novel non-invasive color based intelligent diagnosis model for plaque psoriasis lesion. The system which based on primary color model from images has used artificial neural network (ANN) as the decision model to discriminate plaque from other major psoriasis....
Saved in:
| Main Author: | |
|---|---|
| Format: | Research Reports |
| Language: | en |
| Published: |
2005
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/49558/1/49558.pdf https://ir.uitm.edu.my/id/eprint/49558/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1833068960330809344 |
|---|---|
| author | Jarmin, Roziah |
| author_facet | Jarmin, Roziah |
| author_sort | Jarmin, Roziah |
| building | Tun Abdul Razak Library |
| collection | Institutional Repository |
| content_provider | Universiti Teknologi Mara |
| content_source | UiTM Institutional Repository |
| continent | Asia |
| country | Malaysia |
| description | This work describes the development of a novel non-invasive color based intelligent diagnosis model for plaque psoriasis lesion. The system which based on primary color model from images has used artificial neural network (ANN) as the decision model to discriminate plaque from other major psoriasis. This model known as multi color spectrum ANN was been designed to utilize all three RGB primary components. The optimized model was evaluated and validated through analysis of the performance indicators applied in medical research; sensitivity, specificity, clustering properties and discriminative power of the models by plotting the effects of threshold adjustment on their diagnostic accuracy, error and uncertainty (DA, DE and DU), and the optimum Euclidean Distance (ED) from the ideal point (1,0) in the receiver characteristics operating (ROC) plot. Other than that, the model's network structure was also considered. Findings have showed that the uniqueness of ANN model in recognizing and relating the input-output pattern with no-prior knowledge about this relationship has made the multi color spectrum model to produce reliable dermatological diagnosis. This model, which based only on mean gradation indices (x) of the three primary components (RGB) and reflecting only the location information of the lesion samples data histogram, produced high accuracy (75%) with a specificity (85.71%) and sensitivity of 88.10%. This model on the contrary, has one setback where it consumed large network size. If efficiency is preferred rather than cost, then this optimized model should be selected as the novel non-invasive color based intelligent diagnosis model for plaque psoriasis lesion. Finally, this work has contributed to a possible solution for the application of biomedical imaging in a medical profession. |
| format | Research Reports |
| id | my.uitm.ir-49558 |
| institution | Universiti Teknologi Mara |
| language | en |
| publishDate | 2005 |
| record_format | eprints |
| spelling | my.uitm.ir-495582021-08-18T02:24:30Z https://ir.uitm.edu.my/id/eprint/49558/ Application of ANN in discriminating skin lesions / Roziah Jarmin Jarmin, Roziah Biomedical engineering Computer applications to medicine. Medical informatics Neural Networks (Computer). Artificial intelligence This work describes the development of a novel non-invasive color based intelligent diagnosis model for plaque psoriasis lesion. The system which based on primary color model from images has used artificial neural network (ANN) as the decision model to discriminate plaque from other major psoriasis. This model known as multi color spectrum ANN was been designed to utilize all three RGB primary components. The optimized model was evaluated and validated through analysis of the performance indicators applied in medical research; sensitivity, specificity, clustering properties and discriminative power of the models by plotting the effects of threshold adjustment on their diagnostic accuracy, error and uncertainty (DA, DE and DU), and the optimum Euclidean Distance (ED) from the ideal point (1,0) in the receiver characteristics operating (ROC) plot. Other than that, the model's network structure was also considered. Findings have showed that the uniqueness of ANN model in recognizing and relating the input-output pattern with no-prior knowledge about this relationship has made the multi color spectrum model to produce reliable dermatological diagnosis. This model, which based only on mean gradation indices (x) of the three primary components (RGB) and reflecting only the location information of the lesion samples data histogram, produced high accuracy (75%) with a specificity (85.71%) and sensitivity of 88.10%. This model on the contrary, has one setback where it consumed large network size. If efficiency is preferred rather than cost, then this optimized model should be selected as the novel non-invasive color based intelligent diagnosis model for plaque psoriasis lesion. Finally, this work has contributed to a possible solution for the application of biomedical imaging in a medical profession. 2005 Research Reports NonPeerReviewed text en https://ir.uitm.edu.my/id/eprint/49558/1/49558.pdf Application of ANN in discriminating skin lesions / Roziah Jarmin. (2005) [Research Reports] (Unpublished) |
| spellingShingle | Biomedical engineering Computer applications to medicine. Medical informatics Neural Networks (Computer). Artificial intelligence Jarmin, Roziah Application of ANN in discriminating skin lesions / Roziah Jarmin |
| title | Application of ANN in discriminating skin lesions / Roziah Jarmin |
| title_full | Application of ANN in discriminating skin lesions / Roziah Jarmin |
| title_fullStr | Application of ANN in discriminating skin lesions / Roziah Jarmin |
| title_full_unstemmed | Application of ANN in discriminating skin lesions / Roziah Jarmin |
| title_short | Application of ANN in discriminating skin lesions / Roziah Jarmin |
| title_sort | application of ann in discriminating skin lesions / roziah jarmin |
| topic | Biomedical engineering Computer applications to medicine. Medical informatics Neural Networks (Computer). Artificial intelligence |
| url | https://ir.uitm.edu.my/id/eprint/49558/1/49558.pdf https://ir.uitm.edu.my/id/eprint/49558/ |
| url_provider | http://ir.uitm.edu.my/ |