Hakan Temiz

Rapidly increasing volumes of visual content on online travel platforms require hotel photographs not only to be classified accurately but also to be presented to users within a meaningful information organization structure. Accordingly, the effectiveness of AI-enabled visual content management depends not only on the classification model used but also on the category structure adopted by the platform. This study jointly examines model selection and information organization design in the automated management of hotel images. Seven deep learning architectures were compared on a balanced dataset of 7000 hotel images under two alternative category structures, one including and one excluding the Amenities category. The models were evaluated using several performance metrics, primarily accuracy, AUC, and F1-score; class confusions and semantic overlaps were also examined through error analysis. ViT achieved the highest overall performance under both structures and reached an accuracy of 0.961, an AUC of 0.993, and an F1-score of 0.961 in the six-category structure. In the seven-category structure, which included the Amenities category, performance decreased across all models. Error analysis indicated that this decline was particularly associated with visually and semantically overlapping category boundaries. The findings indicate that, when designing AI-enabled content management for online platforms, the classification model and the information organization structure should not be evaluated independently.

Regional disparities in access to medical diagnostic imaging technologies (MDITs) present a significant barrier to achieving health equity, particularly in developing countries. Understanding how these technologies are distributed and utilized is essential for informing equitable health policy. This study examines the distribution and utilization of MDITs across Türkiye’s 12 NUTS regions using a hierarchical clustering approach. Unlike previous studies, the analysis incorporates both technological capacity and utilization (CaU) variables, evaluated jointly and independently. Imaging modalities are also stratified based on their technological complexity and investment requirements to capture nuanced regional patterns. Findings indicate that although Türkiye demonstrates an overall balanced distribution of MDITs, notable regional disparities in utilization efficiency remain. Regions exhibiting similar usage patterns tend to cluster together irrespective of geographic proximity. Interestingly, the clusters often transcend geographical proximity; regions located at opposite ends of the country tend to cluster on the basis of similar utilization patterns. This may suggest that disparities between administrative centers and rural areas are less pronounced than previously assumed. These patterns imply that institutional capacity, healthcare workforce distribution, and demographic demand may have a stronger influence on utilization than spatial location. The study highlights a disconnect between capacity and actual use of diagnostic imaging technologies, underscoring the need for targeted policy interventions. It also suggests that regional utilization patterns may align more with functional similarities than with geographic proximity. Moreover, analyzing technological capacity and utilization variables separately—rather than as a combined index—yielded more transparent and objective insights into regional disparities. These findings contribute to optimizing health resource allocation and support evidence-based policymaking aimed at advancing equitable access to diagnostic services, aligning with Türkiye’s commitment to universal health coverage.

In organizations, employees work with information stored in files according to their duties and responsibilities. Windows uses resource-based access permissions that any permission for any user has to be set separately per resource. This approach gets complicated as the number of resources and users increase, and causes oversights in assigning permissions. Therefore, a special mechanism is required to scrutinize what permissions any employee has on any set of resources. This requirement is circumvented by reversing the Windows approach in terms of user-accessible resources. This approach is implemented by a program allowing quick and easy examination of any type of permissions granted or denied to active directory users on any folder. In this way, administrators can make sure there is no any missing or overlooked setting that could cause a security vulnerability. This approach can easily be extended to scrutinize other resources, and for other local or active directory objects.

Hotel bathrooms are one of the most important places in terms of customer satisfaction, and where the most complaints are reported. To share their experiences, guests rate hotels, comment, and share images of their positive or negative ratings. An important part of the room images shared by guests is related to bathrooms. Guests tend to prove their satisfaction or dissatisfaction with the bathrooms with images in their comments. These Positive or negative comments and visuals potentially affect the prospective guests. In this study, two different versions of a deep learning algorithm were designed to classify hotel bathrooms as satisfactory (good) or unsatisfactory (bad, when any defects such as dirtiness, deficiencies, malfunctions were present) by analyzing images. The best-performer between the two models was determined as a result of a series of extensive experimental studies. The models were trained for each of 144 combinations of 5 hyper-parameter sets with a data set containing more than 11 thousand bathroom images, specially created for this study. The “HotelBath” data set was shared also with the community with this study. Four different image sizes were taken into consideration: 128, 256, 512 and 1024 pixels in both directions. The classification performances of the models were measured with several metrics. Both algorithms showed very attractive performances even with many combinations of hyper-parameters. They can classify bathroom images with very high accuracy. Suh that the top algorithm achieved an accuracy of 92.4% and an AUC (area under the curve) score of 0.967. In addition, other metrics also proved the success of the algorithm. The proposed method can allow the rapid, accurate and automatic detection of such undesired circumstances in hotel bathrooms from images. Such a detection system can allow hotel management to take necessary actions quickly to remedy such unsatisfactory cases.

In organizations, employees work with information stored in files according to their duties and responsibilities. Windows uses resource-based access permissions that any permission for any user has to be set separately per resource. This approach gets complicated as the number of resources and users increase, and causes oversights in assigning permissions. Therefore, a special mechanism is required to scrutinize what permissions any employee has on any set of resources. This requirement is circumvented by reversing the Windows’ approach in terms of user-accessible resources. This approach is implemented by a program allowing quick and easy examination of any type of permissions granted or denied to active directory users on any folder. In this way, administrators can make sure there is no any missing or overlooked setting that could cause a security vulnerability. This approach can easily be extended to scrutinize other resources, and for other local or active directory objects.

Images cannot always be expected to come in a certain standard format and orientation. Deep networks need to be trained to take into account unexpected variations in orientation or format. For this purpose, training data should be enriched to include different conditions. In this study, the effects of data enrichment on the performance of deep networks in the super resolution problem were investigated experimentally. A total of six basic image transformations were used for the enrichment procedures. In the experiments, two deep network models were trained with variants of the ILSVRC2012 dataset enriched by these six image transformation processes. Considering a single image transformation, it has been observed that the data enriched with 180 degree rotation provides the best results. The most unsuccessful result was obtained when the models were trained on the enriched data generated by the flip upside down process. Models scored highest when trained with a mix of all transformations.

Ultrasound offers a safe, non-invasive, and inexpensive way of imaging. However, due to its natural intrinsic imaging characteristics, it produces poor quality images with low resolution (LR) compared to other medical imaging modalities. Various image enhancement techniques have been extensively studied to overcome these shortcomings. Super-resolution (SR) is one of these methods, which endeavor to obtain high resolution (HR) images from LR images while enlarging them. Numerous studies have already utilized different SR techniques in various stages of ultrasound imaging (USI). Unlike other studies, which aimed at obtaining SR in the pre-processing phase or early stages of the post-processing phase of USI, we achieved SR on B-mode ultrasound images, which is the last stage of USI. We constructed a deep convolutional neural network (CNN) and trained it with a very large dataset of B-mode ultrasound images for the scale factors 2, 3, 4, and 8. We evaluated the performance of our proposed model quantitatively with eight image quality measures. The quantitative results revealed that our algorithm is much more successful than other methods at each magnification factor. Furthermore, we also verified that there is a statistically significant difference between our approach and others. Besides, qualitative analysis of the reconstructed images also confirms that it produces much better quality HR images than other methods in terms of the human visual system.