Kevin W. Bowyer - Edge Detector Evaluation

  • Edge Detector Evaluation Using Empirical ROC Curves,
    Kevin W. Bowyer, Christine Kranenburg, and Sean Dougherty.
    Computer Vision and Image Understanding 84 (1), October 2001, 77-103.
    pdf of this paper.
    This paper focuses on evaluating the performance of edge detection algorithms. ... Performance is summarized using receiver operating characteristic curves. ... No other work uses adaptive parameter sampling to construct empirical ROC curves for pixel-level performance evaluation, or has used so many real images, or has compared such a broad selection of detectors.

  • Comparison of Edge Detector Performance Through Use In An Object Recognition Task,
    Min C. Shin, Dmitry B. Goldgof and Kevin W. Bowyer.
    Computer Vision and Image Understanding 84 (1), October 2001, 160-178.
    DOI link.
    Edge detector performance is measured using a particular edge-based object recognition algorithm as a higher-level task. A detector's performance is ranked according to the object recognition performance that it generates. We have used a challenging train and test dataset containing 110 images of jeep-like objects. Six edge detectors are compared and results suggest that (1) the SUSAN edge detector performs best and (2) the ranking of various edge detectors is different from that found in other evaluations.

  • Comparison of Edge Detection Algorithms Using a Structure From Motion Task,
    Min C. Shin, Dmitry B. Goldgof, Kevin W. Bowyer and S. Nikiforou,
    IEEE Transactions on Systems, Man and Cybernetics - Part B 31 (4), August 2001, 589-601.
    DOI link.
    We use the task of structure from motion as a "black box" through which to evaluate the performance of edge detection algorithms. Edge detector goodness is measured by how accurately the SFM could recover the known structure and motion from the edge detection of the image sequences. We use a variety of real image sequences with ground truth to evaluate eight different edge detectors from the literature. Our results suggest that ratings of edge detector performance based on pixel-level metrics and on the SFM are well correlated and that detectors such as the Canny detector and the Heitger detector offer the best performance.

  • Overview of Work In Empirical Evaluation of Computer Vision Algorithms,
    Kevin W. Bowyer and P. Jonathon Phillips,
    in Empirical Evaluation Techniques in Computer Vision, K.W. Bowyer and P.J. Phillips, editors, IEEE Computer Society Press, 1998, pages 1-11.
    pdf version of this chapter.

  • Are Edges Sufficient for Object Recognition?,
    Thomas A. Sanocki, Kevin W. Bowyer, Michael Heath, and Sudeep Sarkar,
    Journal of Experimental Psychology: Human Perception and Performance 24 (1), 340-349, January 1998.
    pdf of this paper.
    The authors argue that the concept of `edges' as used in current research on object recognition obscures the significant difficulties involved in interpreting stimulus information. ... With 1-s exposures, the accuracy of identifying objects in the edge images was found to be less than half that with color photographs. Therefore, edges are far from being sufficient for object recognition.

  • Comparison of Edge Detectors: A Methodology and Initial Study,
    Michael Heath, Sudeep Sarkar, Thomas A. Sanocki and Kevin W. Bowyer,
    Computer Vision and Image Understanding 69, (1), 38-54, January 1998.
    DOI link.
    Because of the difficulty of obtaining ground truth for real images, the traditional technique for comparing low-level vision algorithms is to present image results, side by side, and to let the reader subjectively judge the quality. This is not a scientifically satisfactory strategy. However, human rating experiments can be done in a more rigorous manner to provide useful quantitative conclusions. We present a paradigm based on experimental psychology and statistics, in which humans rate the output of low level vision algorithms. We demonstrate the proposed experimental strategy by comparing four well-known edge detectors: ...

  • A Robust Visual Method for Assessing the Relative Performance of Edge-Detection Algorithms,
    Michael Heath, Sudeep Sarkar, Thomas A. Sanocki, and Kevin W. Bowyer,
    IEEE Transactions on Pattern Analysis and Machine Intelligence 19 (12), 1338-1359, December 1997.
    pdf of this paper.
    A new method for evaluating edge detection is presented ... The basic measure of performance is a visual rating score which indicates the perceived quality of the edges for identifying an object. ... The novel aspect of this work is the use of a visual task and real images of complex scenes in evaluating edge detectors.