Abdullah Bulbul
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
Abdullah Bulbul has a total of 34 co-authors. Here are the top co-authors by number of shared publications:
Tolga Capin
TED Univ, Dept Comp Engn, Ankara, Turkey
11 shared publications- Zeynep Cipiloglu Yildiz
Department of Computer Engineering, Faculty of Engineering and Natural Sciences, Manisa Celal Bayar University, Yunusemre, Manisa 45140, Türkiye
5 shared publications Bulent Ozguc
Bilkent University, Turkey
2 shared publications- Rozenn Dahyot
Maynooth University
2 shared publications Sami Arpa
Bilkent University, Turkey
2 shared publications- Semihanur Aktay
Computer Engineering TED University Ankara Turkey
2 shared publications
Latest Publications
- 4 March 2026
- Published by Wiley in Computer Animation and Virtual Worlds
- Vol. 37 (2) , e70108
Perceptual Control of Difficulty for Visual‐Search‐Based Games in 3D Virtual Environments
The arrangement of objects in a 3D gaming environment is key to guiding players' visual attention. The human visual system is naturally drawn to salient areas, characterized by high contrast and importance, which directly affect performance in games like hidden‐object and first‐person shooters. This study investigates how improving user experience can be achieved by varying game difficulty according to objects' saliency. We created two games with three distinct levels of difficulty in which players utilized virtual reality to locate targets in various environments. Game difficulty was modulated by the saliency of these objects within their environments. In the second experiment, we applied the same structure but focused on how varying the materials and colors of objects influenced detection. Our findings show that controlling object saliency and visual characteristics can effectively manage game difficulty, offering new possibilities for adaptive and immersive gaming experiences.
- 4 March 2026
- Published by Wiley in Computer Animation and Virtual Worlds
- Vol. 37 (2) , e70108
Perceptual Control of Difficulty for Visual‐Search‐Based Games in 3D Virtual Environments
The arrangement of objects in a 3D gaming environment is key to guiding players' visual attention. The human visual system is naturally drawn to salient areas, characterized by high contrast and importance, which directly affect performance in games like hidden‐object and first‐person shooters. This study investigates how improving user experience can be achieved by varying game difficulty according to objects' saliency. We created two games with three distinct levels of difficulty in which players utilized virtual reality to locate targets in various environments. Game difficulty was modulated by the saliency of these objects within their environments. In the second experiment, we applied the same structure but focused on how varying the materials and colors of objects influenced detection. Our findings show that controlling object saliency and visual characteristics can effectively manage game difficulty, offering new possibilities for adaptive and immersive gaming experiences.
- Open Access
- 14 January 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (2) , 860
FacadeGAN: Facade Texture Placement with GANs
This study presents a texture-aware image synthesis framework designed to generate material-consistent façades using adversarial learning. The proposed architecture incorporates a mask-guided channel-wise attention mechanism that adaptively merges segmentation information with texture statistics to reconcile structural guiding with textural fidelity. A thorough comparative analysis was performed utilizing three internal variants—Vanilla GAN, Wasserstein GAN (WGAN), and WGAN-GP—against leading baselines, including TextureGAN and Pix2Pix. The assessment utilized a comprehensive multi-metric framework that included SSIM, FID, KID, LPIPS, and DISTS, in conjunction with a VGG-19 based perceptual loss. Experimental results indicate a notable divergence between pixel-wise accuracy and perceptual realism; although established baselines attained elevated PSNR values, the suggested Vanilla GAN and WGAN models exhibited enhanced perceptual fidelity, achieving the lowest LPIPS and DISTS scores. The WGAN-GP model, although theoretically stable, produced smoother but less complex textures due to the regularization enforced by the gradient penalty term. Ablation investigations further validated that the attention mechanism consistently enhanced structural alignment and texture sharpness across all topologies. Thus, the study suggests that Vanilla GAN and WGAN architectures, enhanced by attention-based fusion, offer an optimal balance between realism and structural fidelity for high-frequency texture creation applications.
- 14 January 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (2) , 860
Open AccessFacadeGAN: Facade Texture Placement with GANs
This study presents a texture-aware image synthesis framework designed to generate material-consistent façades using adversarial learning. The proposed architecture incorporates a mask-guided channel-wise attention mechanism that adaptively merges segmentation information with texture statistics to reconcile structural guiding with textural fidelity. A thorough comparative analysis was performed utilizing three internal variants—Vanilla GAN, Wasserstein GAN (WGAN), and WGAN-GP—against leading baselines, including TextureGAN and Pix2Pix. The assessment utilized a comprehensive multi-metric framework that included SSIM, FID, KID, LPIPS, and DISTS, in conjunction with a VGG-19 based perceptual loss. Experimental results indicate a notable divergence between pixel-wise accuracy and perceptual realism; although established baselines attained elevated PSNR values, the suggested Vanilla GAN and WGAN models exhibited enhanced perceptual fidelity, achieving the lowest LPIPS and DISTS scores. The WGAN-GP model, although theoretically stable, produced smoother but less complex textures due to the regularization enforced by the gradient penalty term. Ablation investigations further validated that the attention mechanism consistently enhanced structural alignment and texture sharpness across all topologies. Thus, the study suggests that Vanilla GAN and WGAN architectures, enhanced by attention-based fusion, offer an optimal balance between realism and structural fidelity for high-frequency texture creation applications.
- 30 September 2025
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 42 (10) , 6844-6869
Virtual Reality for Cultural Heritage: A Systematic Review and the Proposal of the AMUX-VR Framework
- 30 September 2025
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 42 (10) , 6844-6869
Virtual Reality for Cultural Heritage: A Systematic Review and the Proposal of the AMUX-VR Framework
- Open Access
- 28 January 2025
- Published by Pensoft Publishers in JUCS - Journal of Universal Computer Science
- Vol. 31 (1) , 3-21
Computational Game Unit Balancing based on Game Theory
Optimizing game elements through iterative human playtests can be time-consuming and insufficient for games with complex intransitive mechanics. Imbalances in games often require the release of numerous balance patches. We present a computational method for making each game unit equally preferable against a uniform play strategy. We leverage concepts from game theory to model intricate relationships among intransitive entities. Matching units against each other is modeled as a symmetric zero-sum game, where unit selection represents a strategy and the error is quantified using payoff values derived from unit parameters. The algorithm takes the initial unit parameters provided by the game designer and optimizes them with minimal changes using gradient descent. Consequently, the payoff matrix converges to a state where a uniform strategy is a near Nash equilibrium, ensuring that each unit is equally preferable under the optimized condition. We implemented a testing environment based on fictitious play and verified our results on different scenarios. While the majority of game theory research focuses on finding optimal strategies given specific environmental conditions, this paper takes a different perspective within the context of game design. We explore game theoretic concepts to address the goal of designing environments that lead to desired strategy choices.
- 28 January 2025
- Published by Pensoft Publishers in JUCS - Journal of Universal Computer Science
- Vol. 31 (1) , 3-21
Open AccessComputational Game Unit Balancing based on Game Theory
Optimizing game elements through iterative human playtests can be time-consuming and insufficient for games with complex intransitive mechanics. Imbalances in games often require the release of numerous balance patches. We present a computational method for making each game unit equally preferable against a uniform play strategy. We leverage concepts from game theory to model intricate relationships among intransitive entities. Matching units against each other is modeled as a symmetric zero-sum game, where unit selection represents a strategy and the error is quantified using payoff values derived from unit parameters. The algorithm takes the initial unit parameters provided by the game designer and optimizes them with minimal changes using gradient descent. Consequently, the payoff matrix converges to a state where a uniform strategy is a near Nash equilibrium, ensuring that each unit is equally preferable under the optimized condition. We implemented a testing environment based on fictitious play and verified our results on different scenarios. While the majority of game theory research focuses on finding optimal strategies given specific environmental conditions, this paper takes a different perspective within the context of game design. We explore game theoretic concepts to address the goal of designing environments that lead to desired strategy choices.
- Open Access
- 29 September 2023
- Published by The Scientific and Technological Research Council of Turkey (TUBITAK-ULAKBIM) - DIGITAL COMMONS JOURNALS in Turkish Journal of Electrical Engineering and Computer Sciences
- Vol. 31 (5) , 892-907
A machine learning approach for dyslexia detection using Turkish audio records
Dyslexia is a learning disorder, characterized by impairment in the ability to read, spell, and decode letters. It is vital to detect dyslexia in earlier stages to reduce its effects. However, diagnosing dyslexia is a time-consuming and costly process. In this paper, we propose a machine-learning model that predicts whether a Turkish-speaking child has dyslexia using his/her audio records. Therefore, our model can be easily used by smart phones and work as a warning system such that children who are likely to be dyslexic according to our model can seek an examination by experts. In order to train and evaluate, we first create a unique dataset that includes audio recordings of 12 dyslexic children and 13 nondyslexic children in an 8-month period. We explore various machine learning algorithms such as KNN and SVM and use the following features: Mel-frequency cepstral coefficients, reading rate, reading accuracy, the ratio of missing words, and confidence scores of the speech-to-text process. In our experiments, we show that children with dyslexia can be detected with 95.63% accuracy even though we use single-sentence long audio records. In addition, we show that the prediction performance of our model is similar to that of the humans?. In this paper, we provide a preliminary study showing that detecting children with dyslexia based on their audio records is possible. Once the mobile application version of our model is developed, parents can easily check whether their children are likely to be dyslexic or not, and seek professional help accordingly.
- 29 September 2023
- Published by The Scientific and Technological Research Council of Turkey (TUBITAK-ULAKBIM) - DIGITAL COMMONS JOURNALS in Turkish Journal of Electrical Engineering and Computer Sciences
- Vol. 31 (5) , 892-907
Open AccessA machine learning approach for dyslexia detection using Turkish audio records
Dyslexia is a learning disorder, characterized by impairment in the ability to read, spell, and decode letters. It is vital to detect dyslexia in earlier stages to reduce its effects. However, diagnosing dyslexia is a time-consuming and costly process. In this paper, we propose a machine-learning model that predicts whether a Turkish-speaking child has dyslexia using his/her audio records. Therefore, our model can be easily used by smart phones and work as a warning system such that children who are likely to be dyslexic according to our model can seek an examination by experts. In order to train and evaluate, we first create a unique dataset that includes audio recordings of 12 dyslexic children and 13 nondyslexic children in an 8-month period. We explore various machine learning algorithms such as KNN and SVM and use the following features: Mel-frequency cepstral coefficients, reading rate, reading accuracy, the ratio of missing words, and confidence scores of the speech-to-text process. In our experiments, we show that children with dyslexia can be detected with 95.63% accuracy even though we use single-sentence long audio records. In addition, we show that the prediction performance of our model is similar to that of the humans?. In this paper, we provide a preliminary study showing that detecting children with dyslexia based on their audio records is possible. Once the mobile application version of our model is developed, parents can easily check whether their children are likely to be dyslexic or not, and seek professional help accordingly.
- Open Access
- 27 January 2023
- Published by Elsevier in Graphical Models
- Vol. 126
Procedural generation of semantically plausible small-scale towns
- 27 January 2023
- Published by Elsevier in Graphical Models
- Vol. 126
Open AccessProcedural generation of semantically plausible small-scale towns
- Open Access
- 16 October 2021
- Published by Mugla Sitki Kocman University in Mugla Journal of Science and Technology
- Vol. 7, 19-26
A PATH GUIDED METHOD FOR QUICKLY LABELING ROADS IN SATELLITE IMAGES
Determining roads from satellite images has gained more research interest after the recent progress on data-heavy machine learning methods which are also accelerated by the increased amounts of accessible data. An important challenge of learning-based approaches is obtaining labeled data to train the systems. In this study, we propose a method for quickly labeling roads over satellite images of any desired location. Our method exploits the 2D path information obtained from OpenStreetMap, an online community-contributed source of geolocated information. In this environment, roads are roughly described as line segments without their exact shapes and sizes. Using this rough information, we propose a simple interactive user interface where users easily label the road boundaries over presented satellite images. Using our approach, it is possible to rapidly label regions with different road characteristics. Such an approach allows for training separate machine learning systems for different parts of the world which would be advantageous over training a single system to identify all kinds of roads.
- 16 October 2021
- Published by Mugla Sitki Kocman University in Mugla Journal of Science and Technology
- Vol. 7, 19-26
Open AccessA PATH GUIDED METHOD FOR QUICKLY LABELING ROADS IN SATELLITE IMAGES
Determining roads from satellite images has gained more research interest after the recent progress on data-heavy machine learning methods which are also accelerated by the increased amounts of accessible data. An important challenge of learning-based approaches is obtaining labeled data to train the systems. In this study, we propose a method for quickly labeling roads over satellite images of any desired location. Our method exploits the 2D path information obtained from OpenStreetMap, an online community-contributed source of geolocated information. In this environment, roads are roughly described as line segments without their exact shapes and sizes. Using this rough information, we propose a simple interactive user interface where users easily label the road boundaries over presented satellite images. Using our approach, it is possible to rapidly label regions with different road characteristics. Such an approach allows for training separate machine learning systems for different parts of the world which would be advantageous over training a single system to identify all kinds of roads.
- Open Access
- 16 October 2021
- Published by Mugla Sitki Kocman University in Mugla Journal of Science and Technology
- Vol. 7, 27-33
PROCEDURAL 3D FOUNTAIN MODELING
Procedural modeling is a method that determines production rules and allows for developing effective structural design in architecture. Fountain (Shadirvan) is an important component of Turkish architecture and having a symmetrical structure and distinctive elements makes fountains suitable for procedural modeling. In this study, we propose a procedural method to design different fountain models similar to ones that exist in Turkish architecture. We specify the parameters by looking at the texture, shape, and quantity of observed fountain images for production rules. Based on the selection of parameters, different fountain models are generated in a procedural manner. We present several different models, determined parameters, and parameter values. We explain how we create shapes such as polygon and pyramid. We show samples of 3d fountain models that our model can produce. Also, we conducted an experiment to evaluate the proposed system in terms of its expressiveness and usability. In this experiment, we asked subjects to produce 3D fountain models similar to the ones in provided images by selecting parameters of the proposed model or by traditional 3D modeling tools.
- 16 October 2021
- Published by Mugla Sitki Kocman University in Mugla Journal of Science and Technology
- Vol. 7, 27-33
Open AccessPROCEDURAL 3D FOUNTAIN MODELING
Procedural modeling is a method that determines production rules and allows for developing effective structural design in architecture. Fountain (Shadirvan) is an important component of Turkish architecture and having a symmetrical structure and distinctive elements makes fountains suitable for procedural modeling. In this study, we propose a procedural method to design different fountain models similar to ones that exist in Turkish architecture. We specify the parameters by looking at the texture, shape, and quantity of observed fountain images for production rules. Based on the selection of parameters, different fountain models are generated in a procedural manner. We present several different models, determined parameters, and parameter values. We explain how we create shapes such as polygon and pyramid. We show samples of 3d fountain models that our model can produce. Also, we conducted an experiment to evaluate the proposed system in terms of its expressiveness and usability. In this experiment, we asked subjects to produce 3D fountain models similar to the ones in provided images by selecting parameters of the proposed model or by traditional 3D modeling tools.
- Open Access
- 16 October 2021
- Published by Mugla Sitki Kocman University in Mugla Journal of Science and Technology
- Vol. 7, 11-18
SALIENCY BASED ILLUMINATION CONTROL FOR GUIDING USER ATTENTION IN 3D SCENES
Visual attention has a major impact on how we perceive 3D environments and saliency is a component of visual attention expressing how likely a scene or item is to capture our attention due to its apparent features. Saliency relies upon shape, shading, brightness, and other visual attributes of items. The saliency distribution of a visual field is influenced by the illumination of a scene, which has a significant impact on those visual properties. This work aims to control the saliency by manipulating the illumination parameters in a 3D scene. For this reason, given a sensible 3D scene, the light parameters that provide maximum saliency for the point of interest objects are investigated. In other words, we propose a method for task-aware automatic lighting setup. In this paper, 2D renderings of a 3D scene from various perspectives are considered, and the effects are analyzed in terms of saliency distribution under various lighting conditions. Also, for this process, different saliency estimation methods and calculations are investigated and eye tracker based user experiments are conducted to verify the results.
- 16 October 2021
- Published by Mugla Sitki Kocman University in Mugla Journal of Science and Technology
- Vol. 7, 11-18
Open AccessSALIENCY BASED ILLUMINATION CONTROL FOR GUIDING USER ATTENTION IN 3D SCENES
Visual attention has a major impact on how we perceive 3D environments and saliency is a component of visual attention expressing how likely a scene or item is to capture our attention due to its apparent features. Saliency relies upon shape, shading, brightness, and other visual attributes of items. The saliency distribution of a visual field is influenced by the illumination of a scene, which has a significant impact on those visual properties. This work aims to control the saliency by manipulating the illumination parameters in a 3D scene. For this reason, given a sensible 3D scene, the light parameters that provide maximum saliency for the point of interest objects are investigated. In other words, we propose a method for task-aware automatic lighting setup. In this paper, 2D renderings of a 3D scene from various perspectives are considered, and the effects are analyzed in terms of saliency distribution under various lighting conditions. Also, for this process, different saliency estimation methods and calculations are investigated and eye tracker based user experiments are conducted to verify the results.
- 1 February 2021
- Published by Elsevier in Pattern Recognition
- Vol. 110
Goal driven network pruning for object recognition
- 1 February 2021
- Published by Elsevier in Pattern Recognition
- Vol. 110
Goal driven network pruning for object recognition