Accelerated feature detectors for real-time vision based application

Spatial resolution is a very important quality metric to measure digital images. The higher the resolution of the image, the more image details provided. It is projected to convert the low-resolution(LR) image or video stream to high-resolution(HR) ones before they are displayed/used for analysis or...

Full description

Saved in:
Bibliographic Details
Main Author: Biyani, Divesh
Other Authors: Lam Siew Kei
Format: Final Year Project
Language:English
Published: 2017
Subjects:
Online Access:http://hdl.handle.net/10356/70555
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Spatial resolution is a very important quality metric to measure digital images. The higher the resolution of the image, the more image details provided. It is projected to convert the low-resolution(LR) image or video stream to high-resolution(HR) ones before they are displayed/used for analysis or pattern extraction. All these requirements are hopeful to be satisfied in an inexpensive manner by using super-resolution(SR) technique. To do this, different algorithms are to be looked at and in the past decade there have been many image processing algorithms. The best algorithm should be identified that suits the purpose. Recently a lot of interpolation algorithms have been proposed, but many of these are highly computationally expensive and so cannot be used for real-time applications. Apart from being computationally expensive, many of these algorithms are also space expensive, that is they take up a lot of memory and cannot completely reside in the RAM. In view of real-time applications, I would like to identify an algorithm that is both computationally inexpensive and memory efficient and use it for Super-Resolution.