A New Objective Supervised Edge Detection Assessment Using Hysteresis Thresholds

Hasan Abdulrahman (1) , Baptiste Magnier (1) , Philippe Montesinos (1)

1 LGI2P - Laboratoire de Génie Informatique et Ingénierie de Production

Useful for the visual perception of a human, edge detection remains a crucial stage in numerous image processing applications. Therefore, one of the most challenging goals in contour extraction is to operate algorithms that can process visual information as humans need. Hence, to ensure that it is reliable, an edge detection technique needs to be severely assessed before being used it in a computer vision tools. To achieve this task, a supervised evaluation computes a score between a ground truth edge map and a candidate image. Theoretically, by varying the hysteresis thresholds of the thin edges, the minimum score of the measure corresponds to the best edge map, compared to the ground truth. In this study, a new supervised edge map quality measure is proposed, where the minimum score of the measure is associated with an edge map in which the main structures of the desired objects are distinctive.

Mots clés: Edge detection; Supervised evaluation; Hysteresis

Domaines: Traitement des images [eess.IV]

Hasan Abdulrahman, Baptiste Magnier, Philippe Montesinos. A New Objective Supervised Edge Detection Assessment Using Hysteresis Thresholds. ICIAP 2017 - International Conference on Image Analysis and Processing, Sep 2017, Catania, Italy. pp.3-14, ⟨10.1007/978-3-319-70742-6_1⟩. ⟨hal-01940328⟩