Saliency Detection via Graph-Based Manifold Ranking

This repository provides the source code implementation based on the research paper "Saliency Detection via Graph-Based Manifold Ranking" authored by Chuan Yang, Lihe Zhang, Huchuan Lu, Ming-Hsuan Yang, and Xiang Ruan. The codebase offers the following functionalities:

1. Implements Canny edge detection operator to compute binary edge maps, utilizing gradient thresholding and non-maximum suppression techniques for precise edge identification.

2. Executes graph-based manifold ranking algorithm for saliency detection, constructing affinity matrices and solving ranking functions through matrix operations to identify salient image regions.

3. Provides fundamental image processing functions and utility tools, including matrix manipulation routines and color space conversions, to facilitate advanced image analysis and processing workflows.

4. Supports multiple image format reading/writing operations using imread() and imwrite() functions, along with image transformation capabilities including scaling using imresize() and rotation via imrotate() methods.

5. Includes comprehensive example code with commented demonstrations showing parameter configuration and result visualization, enabling users to quickly understand and utilize the source program effectively.

By leveraging this implementation, researchers can efficiently perform graph-based saliency detection, providing a solid foundation for subsequent image analysis and processing tasks. The authors' research contributions significantly advance the field of computer vision and image processing methodologies.