We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Detection of local motion blurred/non-blurred regions in an image.
- Authors
Kapuriya, B R; Pradhan, Debasish; Sharma, Reena
- Abstract
Motion blur of an image is a common phenomenon that occurs while taking a photograph due to the relative movement of the object and an image acquiring device. It is essential to detect this phenomenon of blurring of images in many applications such as information retrieval. This paper proposes a novel local blur detection technique, and it performs better than the existing works. This technique mainly uses Radon transform and Laplacian of Gaussian on the local neighborhood around each pixel to estimate blur information. Additionally, two new weight functions are introduced based on local geodesic distance and local variance. It is shown that these functions play a significant role in segregating blur and non-blurred parts. Simulation results validate the correctness and accuracy by testing the proposed algorithm on some challenging images with similar color information in the foreground and background. Various quantitative performance measures have determined the superiority of the proposed method.
- Subjects
RADON transforms; GEODESIC distance; INFORMATION retrieval; FOURIER transforms; IMAGE stabilization; ALGORITHMS
- Publication
Multimedia Tools & Applications, 2024, Vol 83, Issue 15, p43705
- ISSN
1380-7501
- Publication type
Article
- DOI
10.1007/s11042-023-17340-3