Consistent validation of gray-level thresholding image segmentation algorithms based on machine learning classifiers
- 41 Downloads
We propose a Machine Learning approach for Image Validation (MaLIV) to rank the performances of two or more outputs obtained from different gray-level thresholding image segmentation algorithms. MaLIV utilizes machine learning classifiers to rank automatically the outputs of different segmentation algorithms accounting for both the computational complexity of the validation experiment and for the robustness of its results. The proposed method resorts to subsampling to find Fisher consistent estimates of validity measures obtained from a sample of pixels of extremely-reduced size. To this purpose, subsampling is combined with three alternative approaches: learning curves, asymptotic regression and convergence in probability. Results of experiments involving the validation of five images segmented through thirteen different algorithms are presented.
KeywordsImage validation Subsampling Learning curves Asymptotic regression Convergence in probability Classifiers’ prediction capabilities MaLIV Machine learning
The research activities of Luca Frigau described in this paper have been conducted within the R&D project “Cagliari2020” partially funded by the Italian University and Research Ministry (grant No. MIUR_PON04a2_00381). The research activities of Luca Frigau, Claudio Conversano and Francesco Mola are supported by the Regione Autonoma della Sardegna under the Grant Pacchetti Integrati di Agevolazione Industria, Artigianato e Servizi, PIA – 2013 No. 282/13 and by the Italian University and Research Ministry (Progetto Dipartimenti di Eccellenza 2018–2022).
- Aria M, D’Ambrosio A, Iorio C, Siciliano R, Cozza V (2018) Dynamic recursive tree-based partitioning for malignant melanoma identification in skin lesion dermoscopic images. Stat Pap. https://doi.org/10.1007/s00362-018-0997-x
- Cortes C, Jackel LD, Solla SA, Vapnik V, Denker JS (1994) Learning curves: asymptotic values and rate of convergence. In: Advances in neural information processing systems, pp 327–334Google Scholar
- El-Samie FA (2012) Image restoration. Lap Lambert Academic Publishing GmbH KG, Saarbrucken ISBN 9783847333531Google Scholar
- Glasbey CA (1993) An analysis of histogram-based thresholding algorithms. CVGIP 55(6):532–537Google Scholar
- Lehmann EL (2004) Elements of large-sample theory. Springer, New YorkGoogle Scholar
- Lowe DG (2004) Distinctive image features from scale-invariant keypoints. Int J Comput Vis 60(2):91–110. https://doi.org/10.1023/B:VISI.0000029664.99615.94 ISSN 1573-1405CrossRefGoogle Scholar
- Martin D, Fowlkes C, Tal D, Malik J (2001) A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics. In: Computer vision, 2001. ICCV 2001. Proceedings. Eighth IEEE international conference on. vol 2. IEEE, pp 416–423Google Scholar
- Moré JJ (1978) The levenberg-marquardt algorithm: implementation and theory. In: Numerical analysis. Springer, pp 105–116Google Scholar
- Narkhede HP (2013) Review of image segmentation techniques. Int J Sci Mod Eng 1(8):54–61Google Scholar
- Sonka M, Hlavac V, Boyle R (2014) Image processing, analysis, and machine vision. Cengage LearningGoogle Scholar
- Zou KH, Warfield SK, Aditya Bharatha, Tempany CMC, Kaus Michael R, Haker SJ, Wells WM, Jolesz FA, Ron Kikinis (2004) Statistical validation of image segmentation quality based on a spatial overlap index: scientific reports. Acad Radiol 11(2):178–189. https://doi.org/10.1016/S1076-6332(03)00671-8 CrossRefGoogle Scholar