The authors present the results of a comparative performance study of algorithms for detecting keypoints in digital images. The Harris, good features to track (GFTT), SIFT, SURF, FAST, ORB, BRISK, and the MSER keypoint detectors were tested using two types of images: POV-Ray simulated images and photographs from the Caltech 256 image dataset. They tested the repeatability of detection of the image keypoints for the evaluated detectors for a series of images with one degree rotations from 0 to 180° (3982 images in total). In the evaluation scenario they adopted an original approach in which they did not hold back a single image to be the reference image. They conclude that the most computationally complex detector, i.e. the SIFT performs best under rotation transformation of images. However, the FAST and ORB detectors, while being less computationally demanding, perform almost equally well. Hence, they can be viable choices in image processing tasks for mobile applications.

References

1. 1)
  - 23. Fernandez, A., Ghita, O., Gonzalez, E., et al: ‘Evaluation of robustness against rotation of LBP, CCR and ILBP features in granite texture classification’, Mach. Vis. Appl., 2011, 22, (6), pp. 913–926.
2. 2)
  - 22. Griffin, G., Holub, A.D., Perona, P.: ‘The Caltech 256’, Caltech Technical Report.
3. 3)
  - 24. Kylberg, G., Sintorn, I.-M.: ‘On the influence of interpolation method on rotation invariance in texture recognition’, Eurasip J. Image Video Process., 2016, 17, 1–12.
4. 4)
  - 14. Lee, M.H., Park, I.K.: ‘Performance evaluation of local descriptors for affine invariant region detector’, Computer Vision - ACCV 2014 Workshops, 2015, 9008, pp. 630–643.
5. 5)
  - 21. Heinly, J., Dunn, E., Frahm, J. M.: ‘Comparative evalution of binary features’. 12th European Conf. Computer Vision, Florence, Italy, October 7–13, 2012, pp. 759–773.
6. 6)
  - 5. Lowe, D.G.: ‘Distinctive image features from scale-Invariant keypoints’, Int. J. Comput. Vis., 2004, 60, (2), pp. 91–110.
7. 7)
  - 7. Rosten, E., Drummond, T.: ‘Machine learning for high-Speed corner detection’, Comput. Vis. ECCV 2006, 2006, 1, pp. 430–443.
8. 8)
  - 6. Baya, H., Essa, A., Tuytelaarsb, T., et al: ‘Distinctive image features from scale-Invariant keypoints’, Comput. Vis. Image Underst.., 2008, 110, (3), pp. 346–359.
9. 9)
  - 16. Hartmann, W., Havlena, M., Schindler, K.: ‘Predicting matchability’. IEEE Conf. Computer Vision and Patter Recognition, 2014, pp. 9–16.
10. 10)
  - 13. Schmid, C., Mohr, R., Bauckhage, C.: ‘Evaluation of interest point detectors’, Int. J. Comput. Vis., 2000, 37, (2), pp. 151–172.
11. 11)
  - 12. Mikolajczyk, K., Schmid, C.: ‘A performance evaluation of local descriptors’, IEEE Trans. Pattern Anal. Mach. Intell., 2005, 27, (10), pp. 1615–1630.
12. 12)
  - 8. Rublee, E., Rabaud, V., Konolige, K., et al: ‘ORB: an efficient alternative to SIFT or SURF’, 2011 IEEE Int. Conf. Computer Vision (ICCV), 2011, pp. 2564–2571.
13. 13)
  - 4. Shi, J., Tomasi, C.: ‘Good features to track’, IEEE Computer Society Conf. Computer Vision and Pattern Recognition, 1994, pp. 593–600.
14. 14)
  - 9. Leutenegger, S., Chliand, M., Siegwart, R. Y.: ‘BRISK: Binary robust invariant scalable keypoints’, IEEE Int. Conf. Computer Vision (ICCV), 2011, pp. 2548–2555.
15. 15)
  - 3. Harris, C., Stephens, M.: ‘A combined corner and edge detectors’. Proc. Fourth Alvey Vision Conf., 1988, pp. 147–151.
16. 16)
  - 20. Tuytelaars, T., Mikolajczyk, K.: ‘Local invariant feature detectors: a survey’, Found. Trends Comput. Graph. Vis., 2007, 3, (3), pp. 177–280.
17. 17)
  - 1. Zhao, B., Hu, T., Shen, L.: ‘Visual odometry – a review of approaches’. 2015 IEEE Int. Conf. Information and Automation, 2015, pp. 2569–2573.
18. 18)
  - 19. Megahed, F. M., Camelio, J. A.: ‘Real-time fault detection in manufacturing environments using face recognition techniques’, J. Intell. Manuf., 2012, 23, (3), pp. 393–408.
19. 19)
  - 17. Khan, N., McCane, B., Mills, S.: ‘Better than SIFT?’, Mach. Vis. Appl., 2015, 26, (6), pp. 819–836.
20. 20)
  - 11. Mikolajczyk, K, Tuytelaars, T., Schmid, C., et al: ‘A comparison of affine region detectors’, Int. J. Comput. Vis., 2005, 65, (1–2), pp. 43–72.
21. 21)
  - 10. Matas, J., Chum, O., Urban, M., et al: ‘Robust wide baseline stereo from maximally stable extremal regions’, Image Vis. Comput., 2002, 22, (10), pp. 761–767.
22. 22)
  - 15. Moreels, P., Perona, P.: ‘Evaluation of features detectors and descriptors based on 3D objects’, Int. J. Comput. Vis., 2007, 73, (3), pp. 263–284.
23. 23)
  - 2. Sharma, K., Moon, I.: ‘Improved scale-invariant feature transform feature-matching technique-based object tracking in video sequences via a neural network and kinect sensor’, J. Electron. Imaging, 2013, 22, (3), 033017, doi: 10.1117/1.JEI.22.3.033017.
24. 24)
  - 18. Mualla, F., Scholl, S., Sommerfeldt, B., et al: ‘Unsupervised unstained cell detection by SIFT keypoint clustering and self-labeling algorithm’, Medical Image Computing and Computer-Assisted Intervention – MICCAI 2014, 2014, 18, pp. 377–384.

Unbiased evaluation of keypoint detectors with respect to rotation invariance

References

Related content