https://repository.ju.edu.et//handle/123456789/1214 2026-05-11T16:24:18Z 2026-05-11T16:24:18Z Samuel Gedefa Getachew Mamo Zerihun Olana https://repository.ju.edu.et//handle/123456789/9285 2024-10-07T06:58:26Z 2023-09-01T00:00:00Z

Classification of Cattle skin Disease using Deep Learning approach. Samuel Gedefa; Getachew Mamo; Zerihun Olana Cattle skin diseases are particular kinds of diseases caused by tumors, allergies, parasites, autoimmune; bacteria infectious, and viral diseases and many cattle skin diseases are very dangerous, mostly if not treated at their first stage. These diseases have caused economic losses in different aspects especially in day-to-day usage as they reduce milk yield, leather quality, and performance in draft cattle. This is huge loss of livestock population by a disease that undermines the efforts towards achieving food security and poverty reduction and cattle skin diseases are becoming the most infectious diseases occurring in cattle of all ages and often found in humans and animals. In the contribution of the cattle skin disease classification many researches have been conducted even if they have gap Different kind of cattle skin like cattle’s located only in Ethiopia To address this problem, we propose an approach for cattle skin disease diagnosis by integrating image processing using deep learning. In this study, 3 model have been developed 2 of them are based up on pre-trained models while one of the model have been developed by ourselves from scratch and to develop the classification model, we have collected 444 cattle skin disease image from Guji and Jimma Zones that we have classify in to Circle Worm Skin Disease, Lumpy Skin Disease, Wart Skin Disease and one Normal class. We split the dataset into 80% for training and 20% for testing. Our CNN scored 85.8% test accuracy and InceptionV3 Model obtained 95.6% test accuracy while EffeicenitnetB0 Model classifies the input symptom image with 99.1 % accuracy

2023-09-01T00:00:00Z Mohammedhassen Abamecha Getachew Mamo Mamo Fideno https://repository.ju.edu.et//handle/123456789/9231 2024-03-28T07:50:09Z 2024-03-01T00:00:00Z

Investigating And Developing Lung Diseases Classification Model Using Ensemble Deep Learning Mohammedhassen Abamecha; Getachew Mamo; Mamo Fideno Lung diseases, caused by the COVID-19 pandemic, pose a significant risk to millions of people. Chest X-ray imaging is widely used for diagnosing lung diseases, but accurate diagnosis remains challenging due to shortages of trained radiologists. A computer-aided recognition system has been proposed to minimize errors, using an ensemble of CNNs. In this research, we present convolutional neural network-based ensembles for classifying chest X-ray images into five classes: Pneumonia, Pneumothorax, Tuberculosis (TB), COVID-19, and normal. To minimize misclassification, we combined three procedures: Balance class, Image augmentation techniques with Keras ImageDataGenerator class & using an ensemble model with transfer learning, three separate CNNs—VGG-16, ResNet-50, and MobileNetV2—are combined to create a picture categorization system. The system trained and tested using 7340 chest X-ray images data type from the National Institute of Health chest X-ray repository and Jimma University Medical Center radiology department, significantly reduces manual visual + errors and can serve as a decision support for physicians. We used 80/20 by splitting the data into 80% for training, 10% for tests, and 10% for validation to train each three models namely VGG-16, MobileNetV2, and ResNet-50 and then we trained the concatenate ensemble of the three models. We compared the results with each other and finally compared them with the concatenated ensemble of the three models. As we compared to the state-of-the-art methods the promising classification performance of our proposed method achieved an accuracy of 97.02% meaning that our model achieved 4.29% more accuracy than the benchmark. While the accuracy of MobileNetV2 is 92.05%, VGG16 is 95.73% and ResNet50 is 89.20% so the high accuracy we obtained is by ensemble which is 97.02%. An ensemble of CNNS models, despite higher computational and modeling costs, offers superior performance and robustness in lung disease classification, outperforming individual models and enhancing classification accuracy

2024-03-01T00:00:00Z Amanuel Bekele Elsabet Wedajo Mizanu Zelalem https://repository.ju.edu.et//handle/123456789/9161 2024-02-09T07:37:39Z 2024-01-24T00:00:00Z

Authentication And Classification Of Ethiopian Coffee Beans Using Deep Learning Approach. Amanuel Bekele; Elsabet Wedajo; Mizanu Zelalem This research paper presents a comprehensive study on the authentication and classification of Ethiopian coffee beans using deep learning algorithms. The objective is to develop an accurate and reliable model for identifying and categorizing coffee beans based on their origin and quality. The methodology employed in this study incorporates a literature review of existing approaches and leverages deep learning algorithms, namely VGG-16, InceptionNetResNetv2, DenseNet121, and MobileNetV2. These architectures have demonstrated excellent performance in various image recognition tasks. The models were trained and evaluated on a carefully curated dataset of Ethiopian coffee beans. The primary evaluation metric used in this study is accuracy. The hyper parameters are carefully set, and the Adam optimizer is employed to enhance model performance. This study used 11,242 datasets from multiple sources, including bean characteristics and geographical origin. To ensure reliable results, a well-defined dataset split strategy was employed, with 60% of the dataset allocated for training the models, and the remaining 40% reserved for testing and validation. The 60:40 split ratios adopted in this research demonstrated notable improvements in model performance. By training the models on a larger portion of the dataset, they exhibited enhanced accuracy and generalization abilities. The testing and validation sets played a critical role in evaluating the models' performance. The experimental results indicate that the VGG-16 model achieved an accuracy of 89%, InceptionNetResNetv2 achieved 100% accuracy, DenseNet121 achieved 90% accuracy, and MobileNetV2 achieved an accuracy of 94%. These high accuracy rates showcase the effectiveness of the deep learning approach in accurately authenticating and classifying Ethiopian coffee beans. However, it is important to acknowledge certain limitations, such as the sample size and the need for external validation. Future research endeavors should focus on expanding the dataset and incorporating external validation to enhance the reliability and generalizability of the models. This study contributes to the existing literature by showcasing the potential of deep learning algorithms in the field of coffee bean authentication and classification, paving the way for improved quality control and consumer confidence in Ethiopian coffee products

2024-01-24T00:00:00Z Moges, Seyfu Jifara, Worku Yimer, Dawud https://repository.ju.edu.et//handle/123456789/9143 2024-01-17T06:27:26Z 2023-12-29T00:00:00Z

Integration of Feature Fusion Strategy on EfficientNet for Skin Cancer Detection and Classification Moges, Seyfu; Jifara, Worku; Yimer, Dawud Skin cancer is a disorder that arises from changes in healthy skin cells that give them the ability to become malignant. Due to a rise in predominance over the past ten years, it is currently placed among the top ten malignancies in terms of frequency. Patients who are unaware of skin cancer may not be encouraged to seek medical attention for minor skin discoloration because many people lack the knowledge required to notice it. One can lessen and manage the detrimental effects of skin cancer with an accurate diagnosis and prompt, efficient therapy. Investigating skin cancer lesions can be difficult due to their comparatively similar forms, complex expression of the disease, and susceptibility to subjective diagnosis. The obtained features from the multiple EfficientNet model tiers are combined using a feature fusion approach to solve this challenge. Therefore, in this study, a system was developed that could detect and classify skin cancer lesions into benign and malignant automatically by using a feature fusion strategy and EfficientNet algorithm with a transfer learning method. The image dataset was collected from a public dataset that is available on Kaggle and the total dataset used is 27560 from both classes benign and malignant. Pre-processing the skin lesions, extracting features using a pre-trained EfficientNet, feature concatenation, and classifying using deep learning EfficientNet algorithm are the primary components of this research. The study method was tested and yielded average results of 93.4% accuracy, 92.3% precision, 94.8% recall, 92.1% specificity, and 93.5% f1-score, respectively as well as confusion matrix achieved 1269(92.00%) true positives, 109(8.00%) false positives, 72(5.00%) false negatives, and 1306(95.00%) true negatives

2023-12-29T00:00:00Z