A robust auto-focus measure based on inner energy

Abstract

This paper proposes a robust auto-focus (AF) measure based on inner energy. In general, the inner energy of noise pixels is close to zero because the magnitude of gradient and the direction of the noise pixels are random. Therefore, the inner energy can effectively eliminate the influence of noise on image quality assessment. But the gradients of near edge points are consistent with those of edge points, so the inner energy of edge pixels is relatively large, and the detail information of the image can be highlighted. Experimental results indicate that compared with traditional methods, the proposed method has higher accuracy, fewer local peaks, stronger robustness and better practicability. In particular, the evaluation results are close to the subjective evaluation of the human eyes. These results illustrate that the proposed method can be applied in automatic focusing.

References

[1] HanJ W, KimJ H, LeeH T, KoS J. IEEE Transactions on Consumer Electronics, 2011, 57: 232

[2] LeeS Y, KumarY, ChoJ M, LeeS W, KimS W. IEEE Transactions on Circuits & Systems for Video Technology, 2008, 18: 1237

[3] PertuzS, PuigD, GarciaM A. Pattern Recognition, 2012, 46: 1157

[4] JJ, JL, JP. IEEE Transactions Consumer Electronics, 2011, 57: 1

[5] ZhangX, WuH, MaY. Applied & Computational Harmonic Analysis, 2015, 40: 430

[6] YangG, NelsonB JWavelet-Based Autofocusing and Unsupervised segmentation of Microscopic ImagesInternational IEEE Conference on Intelligent Robots and Systems, 2003, 3: 2143

[7] Huang J T, Shen C H, Phoong S M and Chen H, Robust Measure of Image Focus in the Wavelet Domain, IEEE International Symposium on Intelligent Signal Processing and Communication Systems, 212 (2005).

[8] KongL J, NieP. Journal of Optoelectronics·Laser, 2016, 27: 198

[9] LVH X, ZhaoZ G, GuoY J, WangF C. Journal of Optoelectronics·Laser, 2016, 27: 77

[10] DonohoD L. IEEE Transactions on Information Theory, 1995, 41: 613

[11] GonzalezR C, WoodsR E, MastersB RDigital Image ProcessingPientice Hall, 2008,

[12] WangZ H, WuF C. Chinese Journal of Computers, 2009, 32: 2211

[13] Subbarao M and Tyan J K, Optimal Focus Measure for Passive Autofocusing and Depth-from-Focus, Symposium on Videometrics IV, 89 (1999).