Faculty of Engineering

Permanent URI for this communityhttps://hdl.handle.net/10679/10

Browse

Browsing by Institution Author "ATEŞ, Hasan Fehmi"

Filter results by typing the first few letters
Now showing 1 - 4 of 4
  • Placeholder
    Conference ObjectPublicationMetadata only
    Deep learning based event recognition in aerial imagery
    (IEEE, 2023) Şahin, A. H.; Ateş, Hasan Fehmi; Electrical & Electronics Engineering; ATEŞ, Hasan Fehmi
    In this paper, we investigate event recognition for aerial surveillance. This is a significant task especially when we consider the growing popularity of UAVs. The main purpose of the paper is to detect events both at the clip level in aerial videos and also at the frame level in aerial images. To achieve this goal, novel deep learning models and training techniques are used. In this work, we propose new model architectures to detect events in both image and video domains. The developed models are tested on the ERA dataset. Results show that the proposed models achieve state-of-the-art performance on both single images and aerial video clips of the ERA dataset.
  • Placeholder
    ArticlePublicationMetadata only
    Deep learning-based blind image super-resolution with iterative kernel reconstruction and noise estimation
    (Elsevier, 2023-08) Ateş, Hasan Fehmi; Yildirim, S.; Gunturk, B. K.; Electrical & Electronics Engineering; ATEŞ, Hasan Fehmi
    Blind single image super-resolution (SISR) is a challenging task in image processing due to the ill-posed nature of the inverse problem. Complex degradations present in real life images make it difficult to solve this problem using naïve deep learning approaches, where models are often trained on synthetically generated image pairs. Most of the effort so far has been focused on solving the inverse problem under some constraints, such as for a limited space of blur kernels and/or assuming noise-free input images. Yet, there is a gap in the literature to provide a well-generalized deep learning-based solution that performs well on images with unknown and highly complex degradations. In this paper, we propose IKR-Net (Iterative Kernel Reconstruction Network) for blind SISR. In the proposed approach, kernel and noise estimation and high-resolution image reconstruction are carried out iteratively using dedicated deep models. The iterative refinement provides significant improvement in both the reconstructed image and the estimated blur kernel even for noisy inputs. IKR-Net provides a generalized solution that can handle any type of blur and level of noise in the input low-resolution image. IKR-Net achieves state-of-the-art results in blind SISR, especially for noisy images with motion blur.
  • Placeholder
    Conference ObjectPublicationMetadata only
    Improved homographic adaptation for keypoint generation in cross-spectral registration of thermal and optical imagery
    (SPIE, 2023) Yağmur, İsmail Can; Ateş, Hasan Fehmi; Electrical & Electronics Engineering; ATEŞ, Hasan Fehmi; Yağmur, İsmail Can
    Autonomous navigation is an important area of research for aerial vehicles. Visual odometry and simultaneous localization and mapping algorithms are critical for the three-dimensional understanding of the environment. For that purpose, consistent multi-spectral maps of the environment should be generated. Existing pixel-based image registration methods are accurate but too slow to operate in real-time. Recently deep learning is used to develop feature-based data-driven methods for generating interest points and associated descriptors for registering multi-spectral image pairs. These methods are fast and perform better than existing methods for optical images. However, the results are less convincing for thermal image registration. In this work, we propose an improved multi-spectral homographic adaptation technique to generate highly repeatable ground truth interest points that are invariant across viewpoint changes in both spectra. These interest points are used to train the MultiPoint image registration network. Simulation results show that our improved model outperforms existing techniques for feature-based image alignment of optical and thermal images.
  • Placeholder
    Conference ObjectPublicationMetadata only
    Swin transformer based siamese network for thermal and optical image registration
    (IEEE, 2023) Elsaeidy, M.; Yağmur, İsmail Can; Ateş, Hasan Fehmi; Güntürk, B. K.; Electrical & Electronics Engineering; ATEŞ, Hasan Fehmi; Yağmur, İsmail Can
    The process of multi-modal image registration is fundamental in remote sensing and visual navigation applications. However, existing image registration methods that are designed for single modality images do not provide satisfactory results when applied to multi-modal image registration. In this research, our objective is to achieve highly accurate alignment of both infrared and optical (visible range) images. To accomplish this goal, we explore the effectiveness of the Swin Transformer encoder and cosine loss in enhancing the keypoint-based image registration process. Simulation results show the improvement achieved in multi-modal registration by using a transformer based Siamese network.