Cem Ünsalan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Cem Ünsalan has a total of 45 co-authors. Here are the top co-authors by number of shared publications:
- Berkan Höke
Agsenze Ltd
30 shared publications Eren Atmaca
30 shared publicationsBeril Sirmacek
Independent scientist, Overijssel, The Netherlands
14 shared publications- Kim L. Boyer
Rensselaer Polytechnic Institute
13 shared publications Abdullah H. Ozcan
Department of Electrical and Electronics Engineering, Computer Vision Research Laboratory, Yeditepe University, Istanbul
9 shared publications- Murat İlSever
Yeditepe University
8 shared publications
Latest Publications
- Open Access
Binary Neural Network Implementation for Handwritten Digit Recognition on FPGA
Binary neural networks provide a promising solution for low-power, high-speed inference by replacing expensive floating-point operations with bitwise logic. This makes them well-suited for deployment on resource-constrained platforms such as FPGAs. In this study, we present a fully custom BNN inference accelerator for handwritten digit recognition, implemented entirely in Verilog without the use of high-level synthesis tools. The design targets the Xilinx Artix-7 FPGA and achieves real-time classification at 80\,MHz with low power consumption and predictable timing. Simulation results demonstrate 84\% accuracy on the MNIST test set and highlight the advantages of manual HDL design for transparent, efficient, and flexible BNN deployment in embedded systems. The complete project including training scripts and Verilog source code are available at GitHub repo for reproducibility and future development.
Open AccessBinary Neural Network Implementation for Handwritten Digit Recognition on FPGA
Binary neural networks provide a promising solution for low-power, high-speed inference by replacing expensive floating-point operations with bitwise logic. This makes them well-suited for deployment on resource-constrained platforms such as FPGAs. In this study, we present a fully custom BNN inference accelerator for handwritten digit recognition, implemented entirely in Verilog without the use of high-level synthesis tools. The design targets the Xilinx Artix-7 FPGA and achieves real-time classification at 80\,MHz with low power consumption and predictable timing. Simulation results demonstrate 84\% accuracy on the MNIST test set and highlight the advantages of manual HDL design for transparent, efficient, and flexible BNN deployment in embedded systems. The complete project including training scripts and Verilog source code are available at GitHub repo for reproducibility and future development.
- Open Access
- 4 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 156486-156496
A Generalized Multiple-Valued FPGA Architecture Based on Improved T-Gate Circuit
Binary logic based FPGAs have been successfully used in a wide range of applications. Although these devices offer clear advantages, their programmable routing requires significant wiring resources between logic blocks. This overhead can be mitigated by employing multiple-valued logic instead of binary logic for signal transmission. Hence, multiple-valued FPGAs emerged. These devices are based on multiple-valued look-up tables and flip-flops within their configurable logic blocks. Although this is a valid strategy, it needs separate combinational and sequential elements which result in high power consumption. To overcome this problem, we propose a novel T-gate based architecture. Our architecture has the advantage of merging the look-up table and flip-flops within multiple-valued logic based configurable logic blocks. The proposed architecture is general such that it is applicable to any multiple-valued logic level and results in significant reduction in power consumption. Moreover, the proposed architecture significantly improves robustness against process variations. We compare the proposed architecture against recent studies to evaluate its performance. We show its strengths and weaknesses this way. All simulation files for the proposed architecture are made publicly available in our GitHub repository [https://github.com/ahmetunu/Nary-LUT/]. Hence, the readers can test and extend them in future studies.
- 4 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 156486-156496
Open AccessA Generalized Multiple-Valued FPGA Architecture Based on Improved T-Gate Circuit
Binary logic based FPGAs have been successfully used in a wide range of applications. Although these devices offer clear advantages, their programmable routing requires significant wiring resources between logic blocks. This overhead can be mitigated by employing multiple-valued logic instead of binary logic for signal transmission. Hence, multiple-valued FPGAs emerged. These devices are based on multiple-valued look-up tables and flip-flops within their configurable logic blocks. Although this is a valid strategy, it needs separate combinational and sequential elements which result in high power consumption. To overcome this problem, we propose a novel T-gate based architecture. Our architecture has the advantage of merging the look-up table and flip-flops within multiple-valued logic based configurable logic blocks. The proposed architecture is general such that it is applicable to any multiple-valued logic level and results in significant reduction in power consumption. Moreover, the proposed architecture significantly improves robustness against process variations. We compare the proposed architecture against recent studies to evaluate its performance. We show its strengths and weaknesses this way. All simulation files for the proposed architecture are made publicly available in our GitHub repository [https://github.com/ahmetunu/Nary-LUT/]. Hence, the readers can test and extend them in future studies.
- Open Access
Integrating Traditional and Deep Learning Methods to Detect Tree Crowns in Satellite Images
Global warming, loss of biodiversity, and air pollution are among the most significant problems facing Earth. One of the primary challenges in addressing these issues is the lack of monitoring forests to protect them. To tackle this problem, it is important to leverage remote sensing and computer vision methods to automate monitoring applications. Hence, automatic tree crown detection algorithms emerged based on traditional and deep learning methods. In this study, we first introduce two different tree crown detection methods based on these approaches. Then, we form a novel rule-based approach that integrates these two methods to enhance robustness and accuracy of tree crown detection results. While traditional methods are employed for feature extraction and segmentation of forested areas, deep learning methods are used to detect tree crowns in our method. With the proposed rule-based approach, we post-process these results, aiming to increase the number of detected tree crowns through neighboring trees and localized operations. We compare the obtained results with the proposed method in terms of the number of detected tree crowns and report the advantages, disadvantages, and areas for improvement of the obtained outcomes.
Open AccessIntegrating Traditional and Deep Learning Methods to Detect Tree Crowns in Satellite Images
Global warming, loss of biodiversity, and air pollution are among the most significant problems facing Earth. One of the primary challenges in addressing these issues is the lack of monitoring forests to protect them. To tackle this problem, it is important to leverage remote sensing and computer vision methods to automate monitoring applications. Hence, automatic tree crown detection algorithms emerged based on traditional and deep learning methods. In this study, we first introduce two different tree crown detection methods based on these approaches. Then, we form a novel rule-based approach that integrates these two methods to enhance robustness and accuracy of tree crown detection results. While traditional methods are employed for feature extraction and segmentation of forested areas, deep learning methods are used to detect tree crowns in our method. With the proposed rule-based approach, we post-process these results, aiming to increase the number of detected tree crowns through neighboring trees and localized operations. We compare the obtained results with the proposed method in terms of the number of detected tree crowns and report the advantages, disadvantages, and areas for improvement of the obtained outcomes.
- 2 December 2024
- Published by Springer Nature in Signal, Image and Video Processing
- Vol. 19 (1) , 1-8
Multi-model tree detection in satellite images with weighted boxes fusion
- 2 December 2024
- Published by Springer Nature in Signal, Image and Video Processing
- Vol. 19 (1) , 1-8
Multi-model tree detection in satellite images with weighted boxes fusion
- 13 August 2024
- Published by Springer Nature
- p. 39-72
Data Acquisition from Sensors
- 13 August 2024
- Published by Springer Nature
- p. 319-353
Convolutional Neural Networks
- 13 August 2024
- Published by Springer Nature
- p. 229-258
Fundamentals of Neural Networks
- 13 August 2024
- Published by Springer Nature
- p. 1-6
Introduction