УДК: 612.82

Deciding on the minimum changes in images of a human face under conditions of indeterminacy

Zhukova, O.V., Shelepin, Y.E., Maksimova, V.A., Vasiliev, P.P., Vershinina, E.A., Fokin, V.A., Sokolov, A.V.

Publication in Journal of Optical Technology

For Russian citation (Opticheskii Zhurnal):

Жукова О.В., Шелепин Ю.Е., Максимова В.А., Васильев П.П., Вершинина Е.А., Фокин В.А., Соколов А.В. Принятие решения о минимальных изменениях в изображениях лица человека в условиях неопределенности // Оптический журнал. 2016. Т. 83. № 12. С. 54–61.

Zhukova O.V., Shelepin Yu.E., Maksimova V.A., Vasiliev P.P., Vershinina E.A., Fokin V.A., Sokolov A.V. Deciding on the minimum changes in images of a human face under conditions of indeterminacy [in Russian] // Opticheskii Zhurnal. 2016. V. 83. № 12. P. 54–61.

For citation (Journal of Optical Technology):

O. V. Zhukova, Yu. E. Shelepin, V. A. Maksimova, P. P. Vasil’ev, E. A. Vershinina, V. A. Fokin, and A. V. Sokolov, "Deciding on the minimum changes in images of a human face under conditions of indeterminacy," Journal of Optical Technology. 83(12), 753-759 (2016). https://doi.org/10.1364/JOT.83.000753

Abstract:

The research problem consisted of using the following methods of neuroiconics: digital image processing, functional magnetic-resonance tomography (fMRT) of brain activity, and psychophysical methods to determine under conditions of indeterminacy the minimum information needed to recognize faces with different intensities of the facial expression. Indeterminacy was exemplified at the semantic level by a pattern of weak facial manifestations of emotions that bear a strong emotional burden and at the physical level by noise of various intensities imposed upon the image. Difference-recognition thresholds were established as a result of analyzing this research, and it was shown that there is correlation between the number of altered pixels in the test images and the correct answers of the observers. Thus, the number of altered pixels in the test images of faces being compared is proportional to the correct answers when a decision is being made in the psychophysical studies and to the number of altered voxels in the images being compared in maps of the brain responses from the data of the fMRT studies.

Keywords:

image, face, facial expression, minimum changes, recognition, neural networks, pixels, voxels, fMRT

Acknowledgements:

The research was supported by the Russian Scientific Fund (14-15-00918).

OCIS codes: 100.4996, 170.6960, 330.5020

References:

1. N. N. Krasilnikov, Yu. E. Shelepin, and O. I. Krasilnikova, “Filtering in the human visual system under threshold-observation conditions,” J. Opt. Technol. 66(1), 3–10 (1999) [Opt. Zh. 66(1), 5–14 (1999)].
2. D. Field, “Matched filters, wavelets and the statistics of natural scenes,” J. Opt. Technol. 66(9), 788–796 (1999) [Opt. Zh. 66(9), 25–36 (1999)].
3. A. K. Zherebko and V. R. Lutsiv, “Matched filtering in natural and artificial neural networks,” J. Opt. Technol. 66(9), 822–824 (1999) [Opt. Zh. 66(9), 69–72 (1999)].
4. “FaceReader software” http://www.noldus.com/human‑behavior‑research/products/facereader.
5. F. V. Campbell and Yu. E. Shelepin, “Possibilities of the fovea in distinguishing objects,” Sens. Sist. 4(2), 181–185 (1990).
6. E. V. Logunova and Yu. E. Shelepin, “Study of the role of spatial-frequency filtering of images when evaluating the age and interpreting the emotional expression of faces,” J. Opt. Technol. 82(10), 694–699 (2015) [Opt. Zh. 82(10), 65–71 (2015)].