Yazar "Karakiş, Rukiye" seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • Küçük Resim Yok
    Öğe
    Image Manipulation Detection Using Residual and Dense Connection-based Deep Learning Models
    (Institute of Electrical and Electronics Engineers Inc., 2024) Karakiş, Rukiye; Gürkahraman, Kali; Ünsal, Emre
    Digital image forgery poses a security threat by manipulating personal data shared on open networks in recent years. In this study, the performance of deep learning models with residual or dense connections, with or without transfer learning, for the classification problem of multiple image manipulations with dual JPEG compression, has been investigated. Transfer learning has produced higher and more stable results at different compression rates. The average accuracy values of the proposed model with dense connections were obtained as 0.9920, 0.9953, and 0.9907 for compression quality factors of 75, 85, and 95, respectively. These values were higher compared to similar studies in the literature. © 2024 IEEE.
  • Küçük Resim Yok
    Öğe
    Optimization of Mixed Pooling Using Genetic Algorithm for Convolutional Neural Networks
    (Institute of Electrical and Electronics Engineers Inc., 2024) Gürkahraman, Kali; Karakiş, Rukiye
    Two of the most commonly used pooling methods in Convolutional Neural Networks are maximum and average pooling. The maximum pooling method may lead to the loss or reduction of discriminative features in high-intensity inputs. On the other hand, in the average pooling method, when average values calculated from local regions are of low intensity, it can cause a decrease in the contrast of feature maps. Therefore, it has been observed that the hybrid pooling method, which weights and combines the outputs of both methods, performs better in some applications. In this study, it is demonstrated that the optimal value of the learnable parameter used for weighting may not be found when updated with all model parameters during the training process. To address this issue, the weighting in hybrid pooling is optimized independently from the model using a Genetic Algorithm (GA). In a multi-classification study conducted with the Flowers dataset, the performance of multi-classification obtained through GA optimization was found to be higher compared to the traditionally learned weight model, based on test results. The accuracy values of the model optimized with GA and the models containing traditionally learned weight were found to be 0.8299 and 0.8109, respectively. © 2024 IEEE.