Project Title
Team
- e20346, Samarakoon S.M.P.H., email
- e20350, Sandamali J.P.D.N., email
- e20419, Wakkumbura W.M.S.S., email
- e20439, Wickramasinghe J.M.W.G.R.L., email
- e20456, Dulaj U.P.S.D., email
Table of Contents
Introduction
Shadows can significantly affect image quality and hinder accurate object detection in many computer vision applications. The Real-Time Shadow Removal Project aims to create an efficient algorithm to detect and remove shadows from images and videos in real-time. By improving visual clarity, the project enables better performance in various applications such as surveillance, autonomous navigation, and augmented reality. The goal is to improve the user experience by enhancing image processing capabilities.
Features
- Real-time shadow detection and removal
- Improved object recognition performance
- Applicable in surveillance, AR, autonomous systems, and more
- Optimized for real-time processing with low latency
Methodology
Key Techniques
-
Shadow Detection: The algorithm identifies shadow regions in images based on changes in pixel color, intensity, and texture.
-
Shadow Removal: Detected shadow regions are processed using techniques like inpainting or advanced image editing to fill in the shadowed areas with realistic pixel values.
-
Real-Time Processing: The algorithm is optimized for processing images and videos without introducing significant delays, making it suitable for applications requiring real-time performance.
Approach
- Preprocessing: The algorithm performs an initial step of enhancing image contrast to improve the detection of shadow regions.
- Post-processing: Final enhancements are applied to ensure the shadow removal to make the image is free from noticeable artifacts.
Sample Results
…..