نوع مقاله : مقاله کامل پژوهشی

نویسندگان

1 دانشجوی دکترای مهندسی برق، گروه الکترونیک، دانشکده‌ی مهندسی برق، دانشگاه علم وصنعت ایران، تهران

2 استاد، گروه الکترونیک، دانشکده‌ی مهندسی برق، دانشگاه علم وصنعت ایران، تهران

چکیده

در این مقاله سعی شده است تا با الگو برداری از سامانه‌ی بینایی انسان، یک روش مقاوم و تکرارپذیر برای بازشناسی اشیا ارائه شود. یکی از معروف­ترین مدل­های ارائه شده مبتنی بر بینایی انسان، مدل HMAX می­باشد که عمل‌کرد مناسبی در بازشناسی اشیا از خود نشان داده است. اما تفاوت­هایی نیز بین این مدل و بینایی انسان وجود دارد، به طوری که رویه‌ی مغز به طور کامل مدل نشده است. از جمله نواقص این مدل می­توان به تکرارناپذیری (حتی در شرایط ثابت)، وجود افزونگی بسیار زیاد و در نتیجه حجم بالای محاسبات و کند بودن اشاره کرد. در این مقاله، سعی شده است تا با مدل کردن عمل‌کرد بخش ثانویه‌ی قشر بینایی و اضافه کردن آن به HMAX، مدل کامل­تری از بینایی انسان ارائه گشته و نقاط ضعف مدل HMAX ، پوشش داده شود. بخش اضافه شده، مانند بخش ثانویه‌ی قشر بینایی، با تمرکز روی ویژگی­های سطح بالاتر و انتخاب ویژگی­های متمایزکننده و البته تکرارپذیر، باعث بهبود یافتن عمل‌کرد مدل خواهد شد. بخش اضافه شده، بار محاسباتی بسیار اندکی داشته به طوری که نه‌تنها باعث کند شدن مدل نمی­شود، بلکه با انتخاب ویژگی­های مختصر و مفید، باعث افزایش سرعت نیز خواهد شد. روش پیشنهادی از لحاظ دقت و زمان پردازش با روش استاندارد مقایسه شده و برتری مدل پیشنهادی نشان داده شده است. علاوه بر آن، تاثیر تعداد ویژگی­های استخراج شده و تعداد تصاویر مورد استفاده جهت آموزش، مورد بررسی قرار گرفته است تا برتری روش پیشنهادی، به ویژه در زمانی که تعداد تصاویر اندکی در دست می‌باشد، نشان داده شود.

کلیدواژه‌ها

موضوعات

عنوان مقاله [English]

Feature Extraction for Object Recognition Inspired by Human Visual System

نویسندگان [English]

  • Hiwa Sufikarimi 1
  • Karim Mohammadi 2

1 Ph.D Student, Electronic School, Electrical Engineering Faculty, Iran University of Science and Technology, Tehran, Iran

2 Professor, Electronic School, Electrical Engineering Faculty, Iran University of Science and Technology, Tehran, Iran

چکیده [English]

In this paper, we tried to present a robust and reliable approach to object recognition by inspiring human visual system. A famous model, inspiring mammalian visual system, is HMAX (Hierarchical Model and X). It shows significant accuracy rates on object recognition tasks. However, there are some differences between this model and human visual system. Indeed cortex's functions are not properly modeled. Unrepeatability under fixed conditions, redundancy, high computing load and being slow are some drawbacks of HMAX. By modeling the secondary visual cortex and adding to the HMAX, we tried to introduce a more accurate model of the human visual system and cover the drawbacks of the previous models. The proposed approach functionally mimics the secondary visual cortex. Attending to high-level features, selecting discriminative and repeatable features, it has higher performance than standard HMAX. The added parts have negligible computation load. Therefore, it does not slow down this model. On the contrary, by selecting brief and useful features, the speed of the model is increased. The proposed approach is compared to the standard HMAX in terms of speed and accuracy rate. The results showed the advantage of proposed approach rather than the standard HMAX. In addition, the effect of the number of features and training images on their performance was shown. It is shown that the proposed approach has a better performance than the standard HMAX especially when the number of feature and training images is small.

کلیدواژه‌ها [English]

  • Object Recognition
  • Feature Extraction
  • Biologically Inspired
  • Repeatability
  • HMAX
[1]     H. Bay, A. Ess, T. Tuytelaars, and L. Van Gool, “Speeded-Up Robust Features (SURF),” Comput. Vis. Image Underst., vol. 110, no. 3, pp. 346–359, Jun. 2008.
[2]     D. G. Lowe, “Distinctive Image Features from Scale-Invariant Keypoints,” Int. J. Comput. Vis., vol. 60, no. 2, pp. 91–110, Nov. 2004.
[3]     N. Dalal and B. Triggs, “Histograms of Oriented Gradients for Human Detection,” in 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), 2005, vol. 1, pp. 886–893.
[4]     P. Moreno, M. J. Marín-Jiménez, A. Bernardino, J. Santos-Victor, and N. P. de la Blanca, “A Comparative Study of Local Descriptors for Object Category Recognition: SIFT vs HMAX,” in Pattern Recognition and Image Analysis, no. June, 2007, pp. 515–522.
[5]     P. T. Riesenhuber M, “Hierarchical models of object recognition in cortex,” Nat. Neurosci., pp. 1019–1025, 1999.
[6]     T. Serre, L. Wolf, S. Bileschi, M. Riesenhuber, and T. Poggio, “Robust object recognition with cortex-like mechanisms. TL  - 29,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 29 VN-r, no. 3, pp. 411–426, 2007.
[7]     S. Seifzadeh, M. Rezaei, and O. Farahbakhsh, “A Computational Visual Neuroscience Model for Object Recognition,” J. Adv. Med. Sci. Appl. Technol., vol. 2, no. 4, p. 315, Jan. 2017.
[8]     H.-Z. Zhang, Y.-F. Lu, T.-K. Kang, and M.-T. Lim, “B-HMAX: A fast binary biologically inspired model for object recognition,” Neurocomputing, vol. 218, pp. 242–250, Dec. 2016.
[9]     Yulong Wang, Qingtian Zhang, and Xiaolin Hu, “Distributed sparse HMAX model,” in 2015 Chinese Automation Congress (CAC), 2015, no. 1, pp. 740–745.
[10] Y. Li, W. Wu, B. Zhang, and F. Li, “Enhanced HMAX model with feedforward feature learning for multiclass categorization,” Front. Comput. Neurosci., vol. 9, no. October, pp. 1–14, Oct. 2015.
[11] C. Theriault, N. Thome, and M. Cord, “Extended Coding and Pooling in the HMAX Model,” IEEE Trans. Image Process., vol. 22, no. 2, pp. 764–777, Feb. 2013.
[12] P. Mishra and B. K. Jenkins, “Hierarchical model for object recognition based on natural-stimuli adapted filters,” in 2010 IEEE International Conference on Acoustics, Speech and Signal Processing, 2010, pp. 950–953.
[13] D. B. Walther and C. Koch, “Attention in hierarchical models of object recognition,” in Progress in Brain Research, vol. 165, no. 06, 2007, pp. 57–78.
[14] H. Sufikarimi and K. Mohammadi, “Speed up biological inspired object recognition, HMAX,” in 2017 3rd Iranian Conference on Intelligent Systems and Signal Processing (ICSPIS), 2017, pp. 183–187.
[15] M. Jazlaeiyan and H. S. Shahhoseini, “Optimal Feature Selection in Biologically Inspired Model for Object Recognition Using Mutual Information Maximisation,” Iran. J. Biomed. Eng., vol. 8, no. 4, pp. 371–383, 2015.
[16] M. Ghodrati, S.-M. Khaligh-Razavi, R. Ebrahimpour, K. Rajaei, and M. Pooyan, “How Can Selection of Biologically Inspired Features Improve the Performance of a Robust Object Recognition Model?,” PLoS One, vol. 7, no. 2, p. e32357, Feb. 2012.
[17] J. Mutch and D. G. Lowe, “Object Class Recognition and Localization Using Sparse Features with Limited Receptive Fields,” Int. J. Comput. Vis., vol. 80, no. 1, pp. 45–57, Oct. 2008.
[18] Y. Lu, M. Lim, H. Zhang, and T. Kang, “Enhanced hierarchical model of object recognition based on a novel patch selection method in salient regions,” IET Comput. Vis., vol. 9, no. 5, pp. 663–672, Oct. 2015.
[19] I. Biederman, “Recognition-by-Components: A Theory of Human Image Understanding,” Psychol. Rev., vol. 94, no. 2, pp. 115–147, 1987.
[20] A. Al Maashri, M. DeBole, C.-L. Yu, V. Narayanan, and C. Chakrabarti, “A hardware architecture for accelerating neuromorphic vision algorithms,” in 2011 IEEE Workshop on Signal Processing Systems (SiPS), 2011, pp. 355–360.
[21] Z. Guo and Z. J. Wang, “An Unsupervised Hierarchical Feature Learning Framework for One-Shot Image Recognition,” IEEE Trans. Multimed., vol. 15, no. 3, pp. 621–632, Apr. 2013.
[22] B. Yang, L. Zhou, and Z. Deng, “C-HMAX: Artificial cognitive model inspired by the color vision mechanism of the human brain,” Tsinghua Sci. Technol., vol. 18, no. 1, pp. 51–56, 2013.
[23] T. Serre, L. Wolf, and T. Poggio, “Object Recognition with Features Inspired by Visual Cortex,” in 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), 2007, vol. 2, pp. 994–1000.
[24] K. D. Flemming, “Essential Neuroscience,” Mayo Clin. Proc., vol. 81, no. 10, p. 1409, Oct. 2006.