Abstract
Simulating the human natural visual system is beneficial for understanding brain intelligence and exploiting new aspects of computer vision. Previous studies have proposed many progressive models and experiments for visual tracking; however, only a few consider all factors involved in visual tracking. Improvements in cross-modal sensory fusion, online physical environment training, and leveraging machine learning are required. In this paper, we present a balanced visual tracking study between neuroscience models and deep-learning methods. In our visual tracking framework, we modify the original region proposal network and interconnect binocular R-CNNs with a new region of interest (RoI) model. Ground frame prediction can be implemented by localization fusion from binocular R-CNNs, as well as external sensory information, such as a dense disparity map. In the behavior stage, visual-motor transformation is implemented through the online training of saccades, pursuit, and vergence networks in the real environment. As demonstrated on a robot, our framework can learn tracking skills through online parameter updates using physical data collected from the robot. The framework achieves performance highly similar to human behaviors and better accuracy than recent models. Moreover, using prediction from our ground vision model to guide binocular, RoI pooling can improve the efficiency of object recognition and localization and reduce visual tracking errors by 27% compared with the original network. In conclusion, this study proposed an effective binocular tracking framework that draws inspiration from brain structures. The performance showed improved accuracy and robustness in tracking random moving targets.
Similar content being viewed by others
Data Availability
The data and programs used in this study are available from the corresponding author upon reasonable request.
References
Escobar MJ, Masson GS, Vieville T, et al. Action recognition using a bio-inspired feedforward spiking network. Int J Comput Vis. 2009;82:284.
Akbarinia A, Parraga CA. Feedback and surround modulated boundary detection. Int J Comput Vis. 2018;126:1367–80.
Tsotsos JK. Motion understanding: task-directed attention and representations that link perception with action. Int J Comput Vision. 2001;45:265–80.
Gupta S, Tolani V, Davidson J, et al. Cognitive mapping and planning for visual navigation. Int J Comput Vis. 2020;128:1311–30.
Porr B, Nürenberg B, Wörgötter F. A VLSI-compatible computer vision algorithm for stereoscopic depth analysis in real-time. Int J Comput Vision. 2002;49:39–55.
Yuille AL, Liu C. Deep Nets: what have they ever done for vision? Int J Comput Vis. 2021;129:781–802.
He K, Gkioxari G, Dollár P, et al. Mask R-CNN. Proc IEEE Int Conf Comput Vis. 2017;2961–2969.
Ren S, He K, Girshick R, Sun J. Faster R-CNN: towards real-time object detection with region proposal networks. Adv Neural Inf Process Syst. 2015;28.
Girshick R. Fast R-CNN. Proc IEEE Int Conf Comput Vis. 2015;1440–1448.
Mahler J, Liang J, Niyaz S, Laskey M, Doan R, Liu X, ... Goldberg K. Dex-Net 2.0: deep learning to plan robust grasps with synthetic point clouds and analytic grasp metrics. 2017. arXiv preprint arXiv:1703.09312.
Wang Z, Fey AM. Deep learning with convolutional neural network for objective skill evaluation in robot-assisted surgery. Int J Comput Assist Radiol Surg. 2018;13(12):1959–70.
Mei C, Sibley G, Cummins M, et al. RSLAM: a system for large-scale mapping in constant-time using stereo. Int J Comput Vis. 2011;94:198–214.
Tai L, Paolo G, Liu M. Virtual-to-real deep reinforcement learning: continuous control of mobile robots for mapless navigation. 2017 IEEE/RSJ Int Conf Intell Robots Syst (IROS). 2017;31–36. IEEE.
Hossain S, Lee DJ. Deep learning-based real-time multiple-object detection and tracking from aerial imagery via a flying robot with GPU-based embedded devices. Sensors. 2019;19(15):3371.
Voigtlaender P, Luiten J, Torr PH, Leibe B. Siam R-CNN: visual tracking by re-detection. Proc IEEE/CVF Conf Comput Vis Pattern Recognit. 2020;6578–6588.
Redmon J, Divvala S, Girshick R, et al. You only look once: unified, real-time object detection. Proc IEEE Conf Comput Vis Pattern Recognit. 2016;779–788.
Antonelli M, Gibaldi A, Beuth F, Duran AJ, Canessa A, Chessa M, Sabatini SP. A hierarchical system for a distributed representation of the peripersonal space of a humanoid robot. IEEE Trans Auton Ment Dev. 2014a;6(4):259–73.
Vannucci L, Ambrosano A, Cauli N, Albanese U, Falotico E, Ulbrich S, ... Laschi C. A visual tracking model implemented on the iCub robot as a use case for a novel neurorobotic toolkit integrating brain and physics simulation. In 2015 IEEE-RAS 15th International Conference on Humanoid Robots (Humanoids). 2015;1179–1184. IEEE.
Pietrini P, Furey ML, Ricciardi E, Gobbini MI, Wu WHC, Cohen L, Haxby JV. Beyond sensory images: object-based representation in the human ventral pathway. Proc Natl Acad Sci. 2004;101(15):5658–63.
Yantis S, Schwarzbach J, Serences JT, et al. Transient neural activity in human parietal cortex during spatial attention shifts. Nat Neurosci. 2002;5(10):995–1002.
Sakata H, Kusunoki M. Organization of space perception: neural representation of three-dimensional space in the posterior parietal cortex. Curr Opin Neurobiol. 1992;2(2):170–4.
Peng J, Srikaew A, Wilkes M, Kawamura K, Peters A. An active vision system for mobile robots. In Smc 2000 conference proceedings. 2000 IEEE International Conference On Systems, Man and Cybernetics.’cybernetics Evolving to Systems, Humans, Organizations, and Their Complex Interactions’. 2000;2:1472–1477. IEEE.
Zhang X, Tay ALP. A physical system for binocular vision through saccade generation and vergence control. Cyber Sys: An Int J. 2009;40(6):549–68.
Antonelli M, et al. A hierarchical system for a distributed representation of the peripersonal space of a humanoid robot. IEEE Trans Auton Ment Dev. 2014b;6(4):259–73.
Kyriakoulis N, Gasteratos A, Mouroutsos SG. An adaptive fuzzy system for the control of the vergence angle on a robotic head. J Intell Fuzzy Syst. 2010;21(6):385–94.
Zhang X, Tay LP. A spatial variant approach for vergence control in complex scenes. Image Vis Comput. 2011;29(1):64–77.
Rea F, Sandini G, Metta G. Motor biases in visual attention for a humanoid robot. 2014 IEEE-RAS Int Conf Humanoid Robots. 2014;779–786. IEEE.
Rea F, Sandini G, Metta G. Motor biases in visual attention for a humanoid robot. 2014 IEEE-RAS Int Conf Humanoid Robots. 2014;779 –786. IEEE.
Kowler E. Eye movements: The past 25 years. Vision Res. 2011;51(13):1457–83.
O’Driscoll GA, Wolff ALV, Benkelfat C, Florencio PS, Lal S, Evans AC. Functional neuroanatomy of smooth pursuit and predictive saccades. NeuroReport. 2000;11(6):1335–40.
Orban de Xivry JJ, Lefevre P. Saccades and pursuit: two outcomes of a single sensorimotor process. J Physiol. 2007;584(1):11–23.
McPeek RM, Keller EL. Saccade target selection in the superior colliculus during a visual search task. J Neurophysiol. 2002;88(4):2019–34.
Wang X, van de Weem J, Jonker P. An advanced active vision system imitating human eye movements. 2013 16th Int Conf Adv Robotics (ICAR). 2013;1–6. IEEE.
Falotico E, Zambrano D, Muscolo GG, Marazzato L, Dario P, Laschi C. Implementation of a bio-inspired visual tracking model on the iCub robot. In 19th International Symposium in Robot and Human Interactive Communication. 2010;564–569. IEEE.
Coombs D, Brown C. Real-time binocular smooth pursuit. Int J Comput Vis. 1993;11(2):147–64.
Das S, Ahuja N. Performance analysis of stereo, vergence, and focus as depth cues for active vision. IEEE Trans Pattern Anal Mach Intell. 1995;17(12):1213–9.
Monaco JP, Bovik AC, Cormack LK. Active, foveated, uncalibrated stereovision. Int J Comput Vis. 2009;85(2):192–207.
Mishra A, Aloimonos Y, Fah CL. Active segmentation with fixation. In 2009 IEEE 12th Int Conf Comput Vis. 2009;468–475. IEEE.
Gibaldi A, Vanegas M, Canessa A, Sabatini SP. A portable bio-inspired architecture for efficient robotic vergence control. Int J Comput Vis. 2017;121(2):281–302.
Zhang Z, Sattler T, Scaramuzza D. Reference pose generation for long-term visual localization via learned features and view synthesis. Int J Comput Vis. 2021;129(4):821–44.
Zhou H, Ummenhofer B, Brox T. DeepTAM: deep tracking and mapping with convolutional neural networks. Int J Comput Vis. 2020;128(3):756–69.
Chen H, Li Y, Deng Y, Lin G. CNN-based RGB-D salient object detection: learn, select, and fuse. Int J Comput Vis. 2021;129(7):2076–96.
Yoon JH, Lee CR, Yang MH, et al. Structural constraint data association for online multi-object tracking. Int J Comput Vis. 2019;127:1–21.
Zhong L, Zhang L. A robust monocular 3D object tracking method combining statistical and photometric constraints. Int J Comput Vis. 2019;127(8):973–92.
Sturm J, Engelhard N, Endres F, Burgard W, Cremers D. A benchmark for the evaluation of RGB-D SLAM systems. In 2012 IEEE/RSJ Int Conf Intell Robots Syst. 2012;573–580. IEEE.
Lemaire T, Berger C, Jung IK, Lacroix S. Vision-based slam: stereo and monocular approaches. Int J Comput Vis. 2007;74(3):343–64.
Agarwal S, Snavely N, Simon I, Seitz S, Szeliski R. Building Rome in a day. Proc ICCV. 2009;72–79.
Kuhn A, Hirschmüller H, Scharstein D, Mayer H. A TV prior for high-quality scalable multi-view stereo reconstruction. Int J Comput Vis. 2017;124(1):2–17.
Kim H, Hilton A. 3D scene reconstruction from multiple spherical stereo pairs. Int J Comput Vis. 2013;104(1):94–116.
Liu A, Marschner S, Snavely N. Caliber: Camera localization and calibration using rigidity constraints. Int J Comput Vis. 2016;118(1):1–21.
Ma C, Chen L, Yong J. AU R-CNN: encoding expert prior knowledge into R-CNN for action unit detection. Neurocomputing. 2019;355:35–47.
Zhu Q, Triesch J, Shi BE. Integration of vergence, cyclovergence, and saccades through active efficient coding. In 2020 Joint IEEE 10th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob). 2020;1–6. IEEE.
de La Bourdonnaye F, Teuliere C, Chateau T, Triesch J. Learning of binocular fixations using anomaly detection with deep reinforcement learning. In 2017 Int Joint Conf Neural Netw (IJCNN). 2017;760–767. IEEE.
Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. 2014. arXiv preprint arXiv:1409.1556.
Araújo AF, Antonino VO, Ponce-Guevara KL. Self-organizing subspace clustering for high-dimensional and multi-view data. Neural Netw. 2020;130:253–68.
Du Y, Yuan C, Li B, Hu W, Yang H, Fu Z, Zhao L. Hierarchical nonlinear orthogonal adaptive-subspace self-organizing map based feature extraction for human action recognition. Proc AAAI Conf Artif Intell. 2008;32(1).
Bernardino A, Santos-Victor J. Vergence control for robotic heads using log-polar images. In Proceedings of IEEE/RSJ Int Conf Intell Robots Syst. IROS’96. 1996;3:1264–1271. IEEE.
Hansen M, Sommer G. Active depth estimation with gaze and vergence control using Gabor filters. Proc 13th Int Conf Pattern Recognit. 1996;1:287–291. IEEE.
Pereyra G, Tucker G, Chorowski J, Kaiser Ł, Hinton G. Regularizing neural networks by penalizing confident output distributions. 2017. arXiv preprint arXiv:1701.06548.
Acknowledgements
We would like to thank Dr. Xinhuan Zhou from the Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, for providing technical assistance and proofreading.
Funding
This study was funded by the National Natural Science Foundation of China (grant number 12172287).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Ethics Approval
This article does not contain any studies involving human participants performed by any of the authors.
Conflict of Interest
The authors declare no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary file1 (MP4 38904 KB)
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Wang, J., Meng, X., Xu, H. et al. Neural Modeling and Real-Time Environment Training of Human Binocular Stereo Visual Tracking. Cogn Comput 15, 710–730 (2023). https://doi.org/10.1007/s12559-022-10091-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12559-022-10091-7