top of page
Lab. LGI2P
وزار Ø© التعليم العالي والبØØ« العلمي - العراق
الجامعة التقنية الشمالية
الدكتور Øسن كريم عبد الرØمن
اختصاص معالجة الصور الرقمية وامنية المعلومات
In image processing tasks, edge detection remains a key point in many applications. The boundaries include the most important structures of the image,
and an efficient boundary detection method should create a contour image containing edges at their correct locations with a minimal of misclassfied pixels.
In the following several comparison measures of different edge detection's involving different edge detectors.
Experiment result 2:
​
Below the experimental evaluations report different assessments for five edge detection methods on several real images: Sobel, Canny, Steerable Filters of order 1 (SF1), Steerable Filters of order 5 (SF5), and Half Gaussian Kernels (HK). All these experiments show comparisons between measure.
All the results are normalized and, compared to the ground truth, a score close to 0 indicates a good edge map whereas a score 1 translates a poor segmentation.
​
Experiment result 1:
​
Experiment result 3:
Experiment result 5:
Experiment result 7:
Experiment result 4:
Experiment result 6:
My future work
References
[5] B. Magnier, P. Montesinos, and D. Diep, "Fast anisotropic edge detection using gamma correction in color images," in IEEE ISPA, 2011, pp. 212-217.
[9] J. Canny, "A computational approach to edge detection," IEEE TPAMI, , no. 6,
pp. 679-698, 1986.
[25] W. T. Freeman and E. H. Adelson, "The design and use of steerable lters," IEEE
TPAMI, vol. 13, pp. 891-906, 1991.
[26] M. Jacob and M. Unser, "Design of steerable lters for feature detection using
canny-like criteria," IEEE TPAMI, vol. 26, no. 8, pp. 1007-1019, 2004
​
Images database
The ground truth and the original image in experimental figures
arises from the database available on the following website:
http://figment.csee.usf.edu/edge/roc/.
​
2017
bottom of page