A Computational Approach to Edge Detection
IEEE Transactions on Pattern Analysis and Machine Intelligence
Context dependent edge detection and evaluation
Pattern Recognition
An optimal linear operator for step edge detection
CVGIP: Graphical Models and Image Processing
Three-dimensional computer vision: a geometric viewpoint
Three-dimensional computer vision: a geometric viewpoint
A performance measure for boundary detection algorithms
Computer Vision and Image Understanding
Edges: saliency measures and automatic thresholding
Machine Vision and Applications
Robust Visual Method for Assessing the Relative Performance of Edge-Detection Algorithms
IEEE Transactions on Pattern Analysis and Machine Intelligence
Local Scale Control for Edge Detection and Blur Estimation
IEEE Transactions on Pattern Analysis and Machine Intelligence
Edge representation with fuzzy sets in blurred images
Fuzzy Sets and Systems
Bounds on Shape Recognition Performance
IEEE Transactions on Pattern Analysis and Machine Intelligence
Comparison of Edge Detectors: A Methodology and Initial Study
CVPR '96 Proceedings of the 1996 Conference on Computer Vision and Pattern Recognition (CVPR '96)
Quantitative evaluation of performance through bootstrapping: edge detection
ISCV '95 Proceedings of the International Symposium on Computer Vision
Gradient histogram: Thresholding in a region of interest for edge detection
Image and Vision Computing
Hi-index | 0.00 |
Most edge detection methods have parameters (threshold values or standard deviation of Gaussian operator for smoothing) to be set, and these parameters make much influence on the outputs of the detectors. In this paper we propose an objective parameter evaluation measure. We evaluate parameters based on the edge ambiguity measures of existence, location and formation. The existence and location ambiguity measures are derived from comparing fuzzy memberships of edgeness with detected edges, and the formation ambiguity measure assesses the connectedness and the total number of edge point in an edge image with respect to the image size. The parameters which produce the least ambiguous edges of a detection method for an image are selected as significant ones. No iterative visual interaction or prior knowledge of edges are needed for these evaluation measures. The effectiveness of the measures is demonstrated by applying the ambiguity measures to synthetic and real images.