Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 23 views

    Single Image Super Resolution

    Uploaded by

    Mutia Hanum
    Single image super resolution aims to generate a high resolution image from a single low resolution input image. It utilizes patch redundancy within the image to estimate missing high frequency details not present in the original low resolution image. Traditional methods like interpolation and sharpening are not sufficient to recover these fine details. Modern single image super resolution methods employ deep learning models trained on large datasets of high and low resolution image pairs to learn the mapping between the two domains. The models learn to replace similar low resolution patches with the corresponding high resolution patches to produce a higher resolution output image with enhanced details.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Single Image Super Resolution

    Uploaded by

    Mutia Hanum
    23 views17 pages
    Single image super resolution aims to generate a high resolution image from a single low resolution input image. It utilizes patch redundancy within the image to estimate missing high frequency details not present in the original low resolution image. Traditional methods like interpolation and sharpening are not sufficient to recover these fine details. Modern single image super resolution methods employ deep learning models trained on large datasets of high and low resolution image pairs to learn the mapping between the two domains. The models learn to replace similar low resolution patches with the corresponding high resolution patches to produce a higher resolution output image with enhanced details.

    Original Description:

    Single Image Super Resolution

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Single image super resolution aims to generate a high resolution image from a single low resolution input image. It utilizes patch redundancy within the image to estimate missing high frequency details not present in the original low resolution image. Traditional methods like interpolation and sharpening are not sufficient to recover these fine details. Modern single image super resolution methods employ deep learning models trained on large datasets of high and low resolution image pairs to learn the mapping between the two domains. The models learn to replace similar low resolution patches with the corresponding high resolution patches to produce a higher resolution output image with enhanced details.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    23 views17 pages

    Single Image Super Resolution

    Uploaded by

    Mutia Hanum
    Single image super resolution aims to generate a high resolution image from a single low resolution input image. It utilizes patch redundancy within the image to estimate missing high frequency details not present in the original low resolution image. Traditional methods like interpolation and sharpening are not sufficient to recover these fine details. Modern single image super resolution methods employ deep learning models trained on large datasets of high and low resolution image pairs to learn the mapping between the two domains. The models learn to replace similar low resolution patches with the corresponding high resolution patches to produce a higher resolution output image with enhanced details.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    of 17

    Single Image Super Resolution

    Sayısal Görüntü İşleme


    DR. Burhan Baraklı

    Mutia Hanum
    14Y5008215
    Outline
    • Introduction to Super-Resolution
    • Super Resolution Techniques
     Multi Image Super Resolution
     Single Image Super Resolution
    Why Super Resolution?

    Definition: Create High Resolution visual output from Low


    Resolution visual input.
    What is Super Resolution?
    • Generating high resolution image with more
    resolving power using one or more low
    resolution images.
    • More resolving power more details
    Terminology
    • Low-Resolution (LR): Pixel density within an
    image is small, therefore offering less details.
    • High-Resolution (HR): Pixel density within an
    image is larger, therefore offering more
    details.
    • Super-Resolution (SR): Obtaining a HR image
    from one or multiple LR images .
    Application
    • Medical imaging (CAT, MRI, etc).
    • Satellite imaging
    • Enlarging consumer photographs
    • Video surveillance (Car wash kidnapping).
    • Converting NTSC video content to high
    definition television
    How to increase resolution?

    • Reducing pixel size (Increase the number of pixels per unit


    area)
    • Increase the chip-size (high cost for high precision optics)
    • Super-resolution (cost less than comparable approaches,
    LR imaging systems can be utilized)
    SUPER RESOLUTION TECHNIQUES
    Multi image super resolution
    • Several images of the same scenery
    • Each image will have different information of
    the same scenery
    Single Image Super Resolution
    • Image database with HR/LR image pairs
    • Replace similar LR patches with corresponding
    HR patches
    Single Image SR
    • Traditional resolution enhancement:
     Smoothing (Gaussian, Wiener, and median filters)
     Interpolation (Nearest neighbor, bilinear, bicubic
    and cubic spline ,etc)
     Sharpening by amplifying existing image details (it
    is useful to do, provided noise isn’t amplified)
    • Single-image SR: Estimate missing high-resolution
    detail that isn’t present in the original image, and
    which we can’t make visible by simple sharpening
    Patch Redundancy in a Single Image
    Employing in-scale Patch Redundancy
    Employing Cross-scale Patch Redundancy

    • Build a cascade of decreasing resolution


    images from the LR image.
    • For each patch in the LR image, search for its
    Nearest Neighbour in the even lower
    resolution image.
    • Take the found neighbour’s parent in the
    original LR image and copy it to be the HR
    image.
    Bicubic interpolation Unified single-image SR (x3) Ground truth image
    Bicubic interpolation Unified single-image SR (x3)
    ÇOK TEŞEKKÜR EDERIM

    You might also like