OKECHUKWU, BRIAN G. (2026) DEVELOPMENT OF A CLINICAL DECISION SUPPORT SYSTEM (CDSS) FOR THORACIC DISEASE DETECTION USING A HYBRID DEEP LEARNING MODEL. Other thesis, Godfrey Okoye University, Enugu Nigeria.
|
Text
DEVELOPMENT OF A CLINICAL DECISION SUPPORT SYSTEM (CDSS).pdf Download (2MB) | Preview |
Abstract
Thoracic diseases, especially pneumonia, continue to be one of the leading causes of death and morbidity across the globe, hence the pressing need for proper diagnosis. In this paper, a clinical decision support system is designed and implemented in order to detect thoracic diseases automatically through the development of a hybrid dheer learning model, namely, DenseNet121- EfficientNetB0, and using chest X-ray images. Here, DenseNet121 is used to acquire the details of the images at a local level while EfficientNetB0 is used to learn in-depth knowledge on the higher semantic level through transfer learning. The combination of features obtained from either of the models helps in improving the performance in classification of thoracic disease. Also, the chest images used for the purpose have been sourced from Kaggle Pneumonia dataset and processed through resizing, normalizing and augmenting the images before the learning starts. Further, during learning, Adam optimization method was utilized while accuracy, precision, recall, f1-score, confusion matrix, ROC curve and average precision metrics were used to evaluate the model. The experimental results show that the model developed had accuracy of 97.75%, AUC of 0.9976 and an average precision score of 0.9973, which clearly indicates the effectiveness of the model. The model developed has already been deployed as a web-based clinical decision support and the users are able to upload their chest X-ray images and get diagnosis.
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | Q Science > Q Science (General) |
| Divisions: | Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science |
| Depositing User: | Cynthia Ugwuoti |
| Date Deposited: | 22 Jul 2026 15:48 |
| Last Modified: | 22 Jul 2026 15:48 |
| URI: | http://eprints.gouni.edu.ng/id/eprint/6061 |
Actions (login required)
 |
View Item |
