Yazar "Alam, Muhammad Zeshan" seçeneğine göre listele

Listeleniyor 1 - 10 / 10
  • Küçük Resim
    Öğe
    Analysis of deep learning based path loss prediction from satellite images
    (Institute of Electrical and Electronics Engineers Inc., 2021) Alam, Muhammad Zeshan; Ateş, Hasan Fehmi; Baykaş, Tunçer; Güntürk, Bahadır Kürşat
    Determining the channel model parameters of a wireless communication system, either by measurements or by running electromagnetic propagation simulations, is a time-consuming process. Any rapid deployment of network demands faster determination of at least major channel parameters. In this paper, we investigate the idea of using deep convolutional neural networks and satellite images for channel parameters (i.e., path loss exponent $n$ and shadowing factor sigma) prediction in a cellular network with aerial base stations. Specifically, we investigate the performance dependency of the method on three different factors: height of the transmitter antenna, quantization levels of the channel parameters and architectural design of CNN. The results presented in this paper show a high prediction accuracy of the channel parameters in real-time.
  • Küçük Resim
    Öğe
    Deconvolution based light field extraction from a single image capture
    (IEEE Computer Society, 2018) Alam, Muhammad Zeshan; Güntürk, Bahadır Kürşat
    In this paper, we propose a method to extract light field using a conventional camera from a single image capture. The method involves an offline calibration process, where point spread functions, relating different perspective images captured with a narrow aperture to a central image captured with a wide aperture, are estimated for different depths. During application, light field perspective images are recovered by de-convolving the input image with the set of point spread functions that were estimated in the offline calibration process.
  • Küçük Resim
    Öğe
    Dynamic range and depth of field extension using camera array
    (Institute of Electrical and Electronics Engineers Inc., 2022) Alam, Muhammad Zeshan; Güntürk, Bahadır Kürşat
    Computational imaging aims to extend the capabilities of conventional imaging systems by extracting richer and more meaningful information through multiple captures or by modifying the camera’s sensors and optics. In this paper, a joint focus stacking and high dynamic range (HDR) imaging method is proposed to extend the limited depth of field and dynamic range of a typical camera. The proposed method acquires a set of input images from a camera array with varying exposure and depth of field, and performs focus stacking under exposure diversity. Experimental results show significant improvement in the dynamic range and depth of field of the final focus-stacked HDR image. Additionally, the proposed method effectively handles the ghosting artifacts during the fusion process.
  • Küçük Resim
    Öğe
    High dynamic range imaging using a plenoptic camera
    (Institute of Electrical and Electronics Engineers Inc., 2017) Wahab, Abdullah; Alam, Muhammad Zeshan; Güntürk, Bahadır Kürşat
    Light field (or plenoptic) imaging has become an attractive research field due to its post-capture capabilities, including refocusing, perspective change and depth estimation. Micro-lens array based cameras that recently emerged have made the light field acquisition process a practical task. In this paper, we propose to convert such a plenoptic camera into a high-dynamic range camera through a minor optical modification. The optical modification is an optical mask placed in front of the main lens to increases the vignetting effect, which darkening towards the borders of the image plane due to loss of light. As a result, different parts of the dynamic range are captured with different sub-aperture images of the light field. These sub-aperture images are then fused through photometric registration and optical flow vectors to produce a high-dynamic range image.
  • Küçük Resim
    Öğe
    Hybrid light field imaging for improved spatial resolution and depth range
    (Springer, 2018) Alam, Muhammad Zeshan; Güntürk, Bahadır Kürşat
    Light field imaging involves capturing both angular and spatial distribution of light; it enables new capabilities, such as post-capture digital refocusing, camera aperture adjustment, perspective shift, and depth estimation. Micro-lens array (MLA)-based light field cameras provide a cost-effective approach to light field imaging. There are two main limitations of MLA-based light field cameras: low spatial resolution and narrow baseline. While low spatial resolution limits the general purpose use and applicability of light field cameras, narrow baseline limits the depth estimation range and accuracy. In this paper, we present a hybrid stereo imaging system that includes a light field camera and a regular camera. The hybrid system addresses both spatial resolution and narrow baseline issues of the MLA-based light field cameras while preserving light field imaging capabilities.
  • Küçük Resim
    Öğe
    Hybrid stereo imaging including a light field and a regular camera
    (Institute of Electrical and Electronics Engineers Inc., 2016) Alam, Muhammad Zeshan; Güntürk, Bahadır Kürşat
    Light field imaging involves capturing both angular and spatial distribution of light; it enables new capabilities, such as post-capture digital refocusing, perspective shift, and depth estimation. Micro-lens array (MLA) based light field cameras provide a cost-effective approach to light field imaging. The two major issues with MLA based light field cameras are low spatial resolution and narrow baseline, which limits the depth range and accuracy. In this paper, we present a hybrid stereo imaging system that includes a light field camera and a regular camera. The hybrid system addresses both the spatial resolution and narrow baseline issues of the MLA based light field cameras while preserving all the light field capabilities.
  • Küçük Resim
    Öğe
    Image enhancement through new techniques in computational photography
    (İstanbul Medipol Üniversitesi Fen Bilimleri Enstitüsü, 2019) Alam, Muhammad Zeshan; Güntürk, Bahadır Kürşat
    Quality of a digital image depends on several factors operative during the image formation process, e.g. sensor defects, sensor dynamic range, poor spatial resolution,lens distortion, camera shake and object motion. This study focused on developing new techniques in computational photography for minimizing some degradations in digital images, including blurring, limited depth of field, low dynamic range, and insufficient resolution. A flexible framework is developed for space-variant deblurring using a single degraded image. Coarse PSF estimation of image patches and PSF clustering are performed to identify regions of uniform blur in an image followed by PSF refinement, deconvolution, and fusion. Focus stacking and high dynamic range (HDR) imaging are combined to generate all-in-focus HDR image using multiple exposure and multiple focus images, captured through a camera array. The limited resolution problem is addressed in the context of light field imaging in two different ways: hybrid stereo imaging involving a regular camera and a light field camera and deconvolution based high-resolution light field extraction from a single image capture. All the developed algorithms are tested on real datasets and both qualitative and quantitative comparisons have been made with the state-of-the-art methods to show the superiority of the proposed algorithms.
  • Küçük Resim
    Öğe
    Light field extraction from a conventional camera
    (Elsevier B.V., 2022) Alam, Muhammad Zeshan; Güntürk, Bahadır Kürşat
    This paper presents a single-shot light field (LF) acquisition method using a conventional camera. The proposed method does not require any modification in the camera design and instead involves an offline calibration process. The calibration step involves the estimation of the point spread function (PSF) for different perspectives at all sampled depths. PSF relates perspective image captured with a narrow aperture opening to the central image captured with wider aperture opening containing multiple perspectives of the same scene superimposed on each other. During LF extraction, different perspective images are obtained by deconvolving the wide aperture input image with the set of PSFs estimated in the offline calibration phase. The experimental results demonstrate that the proposed method generates high angular and spatial resolution LF.
  • Küçük Resim
    Öğe
    Space-variant blur kernel estimation and image deblurring through kernel clustering
    (Elsevier B.V., 2019) Alam, Muhammad Zeshan; Qian, Qinchun; Güntürk, Bahadır Kürşat
    This paper presents a space-variant blur kernel estimation and image deblurring framework. For space-variant blur kernel estimation, the input image is divided into small patches, and for each patch, the blur kernel is estimated. The estimated kernels are then grouped to determine different kernel clusters in the image. During clustering, unreliable kernel estimates are eliminated. The blur kernel for each kernel cluster is finally refined using the corresponding image region, which is the union of image patches associated with the kernels in the cluster. For space-variant image deblurring, the entire image is deconvolved with each blur kernel to produce a set of deblurred images. These images are then fused to produce a blur-free image, where the fusion process selects the optimal regions from the set of deblurred images.
  • Küçük Resim Yok
    Öğe
    Spatio-angular resolution trade-off in face recognition
    (2024) Alam, Muhammad Zeshan; Kelowani, Sousso; Elsaeidy, Mohamed
    Ensuring robustness in face recognition systems across various challenging conditions is crucial for their versatility. State-of-the-art methods often incorporate additional information, such as depth, thermal, or angular data, to enhance performance. However, light field-based face recognition approaches that leverage angular information face computational limitations. This paper investigates the fundamental trade-off between spatio-angular resolution in light field representation to achieve improved face recognition performance. By utilizing macro-pixels with varying angular resolutions while maintaining the overall image size, we aim to quantify the impact of angular information at the expense of spatial resolution, while considering computational constraints. Our experimental results demonstrate a notable performance improvement in face recognition systems by increasing the angular resolution, up to a certain extent, at the cost of spatial resolution.