|
Article information
2025 , Volume 30, ¹ 2, p.126-139
Pham C.T., Tran T.T., Dinh H.H., Duong V.C., Phan M.C.
Computed tomography image reconstruction using graph total variation
In this paper, we propose a model that combines deep learning with graph total variation (GTV) for denoising computed tomography images. The GTV model employs two neural networks: one to represent pixel characteristics and construct a graph, and the other to generate the parameters for the model. Integration of the graph with the learned parameters yields highly effective results, as demonstrated in our experimental evaluations. Our findings indicate that the proposed method outperforms other related approaches in terms of performance.
Keywords: convolutional neural network, total variation, graph total variation, computed tomography, image denoising, deep learning
Author(s):
Pham Cong Thang
Dr.
Position: Lecture
Office: Faculty of Information Technology, The University of Danang University of Science and Technology
Address: 550000, Vietnam, Danang, Nguyen Luong Bang st., 54
E-mail: pcthang@dut.udn.vn
Tran Thi Thu Thao
Position: Research Scientist
Office: Faculty of Statistics and Informatics, The University of Danang University of Economics
Address: 550000, Vietnam, Danang, Ngu Hanh Son st., 71
E-mail: thaotran@due.udn.vn
Dinh Huy Hoang
Position: Research Scientist
Office: Faculty of Information Technology, The University of Danang University of Science and Technology
Address: 550000, Vietnam, Danang, Nguyen Luong Bang st., 54
Duong Van Chon
Position: Research Scientist
Office: Faculty of Information Technology, The University of Danang University of Science and Technology
Address: 550000, Vietnam, Danang, Nguyen Luong Bang st., 54
Phan Manh Cuong
Position: Research Scientist
Office: Faculty of Information Technology, The University of Danang University of Science and Technology
Address: 550000, Vietnam, Danang, Nguyen Luong Bang st., 54
Bibliography link:
Pham C.T., Tran T.T., Dinh H.H., Duong V.C., Phan M.C. Computed tomography image reconstruction using graph total variation // Computational technologies. 2025. V. 30. ¹ 2. P. 126-139
|
