Browsing by Author "Akbacak, E."

Now showing 1 - 2 of 2

Citation Count: 0
Ensemble-Based Alzheimer's Disease Classification Using Features Extracted From Hog Descriptor and Pre-Trained Models
(Sakarya University, 2024) Muzoğlu, N.; Akbacak, E.
Alzheimer's Disease is the most common type of dementia and is a progressive, neurodegenerative disease. The disease worsens over time, and the patient becomes bedridden, unable to move or understand what is happening around him. The main concern of medicine is to slow down the progression of the disease for which no treatment has yet been developed. Artificial intelligence studies have achieved significant success in detecting many diseases. In this study, an artificial intelligence-based approach that uses MR images of the early stage of Alzheimer's Disease to detect the disease at an early stage is presented. Initially, a new dataset was created through the application of the fuzzy technique, thereby expanding the feature space. Then, an ensemble learning-based hybrid deep learning model was developed to reduce the misclassification rate for all classes. The features derived from the inception module, residual modules, and histogram of oriented gradients descriptor are subjected to classification through bagging and boosting algorithms. The proposed model has surpassed many state-of-the-art studies by achieving a high success rate of 99.60% in detecting Alzheimer's disease in its early stages. © 2024, Sakarya University. All rights reserved.
Citation Count: 0
Unet3D Based Next Frame Prediction;
(Institute of Electrical and Electronics Engineers Inc., 2024) Akbacak, E.
The concept of next-frame prediction, which is predicting the subsequent frames using historical frames' spatial and temporal properties, is indispensable in computer vision. There are various application of frame prediction such as predicting a future event in autonomous vehicles, predicting patient falls in biomedical engineering, and reducing the amount of data transmitted in video transmission. Deep learning applications in this field are the focus of the most effective methods. Especially CNN-LSTM, Convolutional LSTMs, and GAN-supported deep learning methods are very common. This study proposes the inflated 3D Unet encoder-decoder model, which is not yet used for the next-frame prediction problem. The proposed model predicts both the next frame and the subsequent frames. Experimental results have shown that the proposed method gives better results than CNN-LSTM and Convolutional LSTMs. © 2024 IEEE.