Abstract
Abstract Region covariance is a robust feature descriptor that allows the use of even the simplest image features like intensity and gradient combined to form a well-performing descriptor for regions on the image. Beyond its robustness, it requires many identical heavy computations on different parts of input data which makes it a good candidate for parallel execution. In this manuscript, we present a real-time parallel implementation of the region covariance which, to our best knowledge, is the first in the literature. We experimented against existing implementations and achieved 6 times faster execution time over vectorized CPU parallel implementation that provides necessary speed up for real-time processing. Additionally, we improved the existing integral image calculation method on CUDA, reducing memory usage by 50%, achieving the fastest computation speed compared to exist- ing solutions, and improved the covariance matrix comparison metric by using a distance metric that is lightweight to compute and easy to implement.
Similar content being viewed by others
References
Acharya KA, Babu R, Vadhiyar SS (2018) A real-time implementation of SIFT using GPU. J Real-Time Image Process 14(2):267–277
Artner NM, Ion A, Kropatsch WG (2011) Multi-scale 2D tracking of articulated objects using hierarchical spring systems. Pattern Recogn 44:800–810
Awad AI (2013) Fingerprint local invariant feature extraction on GPU with CUDA. Informatica (Slovenia) 37:279–284
Bailey T, Durrant-Whyte H (2006) Simultaneous localization and mapping (SLAM): Part II. IEEE Robot Autom Mag 13(3):108–117
Bilgic B, Horn BKP, Masaki I (2010) Efficient integral image computation on the GPU. In: Intelligent vehicles symposium (IV). IEEE, pp 528–533
Cakir S, Aytaç T, Yildirim A, Beheshti S, Gerek ÖN, Cetin AE (2013) Salient point region covariance descriptor for target tracking. Opt Eng 52(2):027207
Chang CH, Kehtarnavaz N (2016) Computationally efficient image deblurring using low rank image approximation and its GPU implementation. J Real-Time Image Proc 12(3):567–573
Cheng J, Grossman M, McKercher Ty (2014) Professional CUDA C programming 1st ed., Wrox Press Ltd., Birmingham, UK, UK
Cherian A, Sra S, Banerjee A, Papanikolopoulos N (2011) Efficient similarity search for covariance matrices via the Jensen-Bregman LogDet divergence. In: IEEE international conference on computer vision, ICCV, pp 2399–2406
Coifman B, Beymer D, McLauchlan P, Malik J (1998) A real-time computer vision system for vehicle tracking and traffic surveillance. Trans Res Part C Emerg Technol 6(4):271–288
Fassolda H, Rosner J (2015) A real-time GPU implementation of the sift algorithm for large-scale video analysis tasks. In: Real-time image and video processing, vol. 9400, p 940007
Faulkner H, Shehu E, Szpak ZL, Chojnacki W, Tapamo JR, Dick A, van den Hengel A (2015) A study of the region covariance descriptor: Impact of feature selection and image transformations. In: International conference on digital image computing: techniques and applications (DICTA)
Förstner W, Moonen B (2003) A metric for covariance matrices. In: Geodesy-the challenge of the 3rd millennium. Springer, pp 299–303
Forsyth DA, Ponce J (2002) Computer vision: A modern approach, pp 88–101
Harris M Parallel prefix sum (Scan) with CUDA. NVIDIA CUDA SDK code samples
Kanopoulos N, Vasanthavada N, Baker RL (1988) Design of an image edge detection filter using the Sobel operator. IEEE J Solid State Circ 23 (2):358–367
Karacan L, Erdem E, Erdem A (2013) Structure-preserving image smoothing via region covariances. ACM Trans Graph (TOG) 32(6):1–11
Lu X et al (2019) Adaptive region proposal with channel regularization for robust object tracking. IEEE Trans Circ Syst Video Technol
Lu X, Ni B, Ma C, Yang X (2019) Learning transform-aware attentive network for object tracking. Neurocomputing 349:133–144
Mader K, Reese G (2012) Using covariance matrices as feature descriptors for vehicle detection from a fixed camera. arXiv preprint arXiv:1202.2528
Mehrez A, Morgan AA, Hemayed EE (2018) Speeding up spatiotemporal feature extraction using GPU. J Real-Time Image Proc, pp 1–29
Nguyen H (2007) Gpu Gems 3. Addison-Wesley Professional. https://developer.nvidia.com/gpugems/GPUGems3/gpugems3_ch39.html
Object tracking with RCD serial C++ implementation. https://bitbucket.org/mark.asbach/opencv-covariance-features/src
Overlap data transfer in CUDA. http://devblogs.nvidia.com/how-overlap-data-transfers-cuda-cc/
Parallel reduction. http://developer.download.nvidia.com/compute/cuda/1.1-Beta/x86_website/projects/reduction/doc/reduction.pdf
Paramkusham S, Rao KMM, Rao BVVSNP (2013) Early stage detection of breast cancer using novel image processing techniques, Matlab and Labview implementation. In: 2013 15th International Conference on Advanced Computing Technologies (ICACT), pp 1–5
Porikli F, Tuzel O, Meer P (2006) Region covariance: A fast descriptor for detection and classification. In: European conference on computer vision, pp 589–600
Połap D, Woźniak M (2019) Bacteria shape classification by the use of region covariance and Convolutional Neural Network. In: 2019 International Joint Conference on Neural Networks (IJCNN). IEEE, Budapest Hungary, pp 1–7
Rister B, Wang G, Wu M, Cavallaro JR (2013) A fast and efficient sift detector using the mobile GPU. In: 2013 IEEE international conference on acoustics speech and signal processing, pp 2674–2678
Schmidhuber J (2015) Deep learning in neural networks: An overview. Neural Netw 61:85–117
Songtao L, Ning J, Zhenxing L, Kanghui J (2018) Saliency detection of infrared image based on region covariance and global feature. J Syst Eng Electron 29(3):483–490
Sudipta SN, Frahm JM, Pollefeys M, Genc Y (2011) Feature tracking and matching in video using programmable graphics hardware. Mach Vis Appl 22:202–217
Tak HM, Joo YH (2016) Localization and autonomous navigation using gpu-based sift and virtual force for mobile robots. Trans Korean Inst Electr Eng 65:1738–1745
Tomasi C, Kanade T (1991) Detection and tracking of point. Int J Comput Vis 9:137–154
Viola P, Jones M (2001) Rapid object detection using a boosted cascade of simple features. In: Proceedings of the 2001 IEEE computer society conference on computer vision and pattern recognition . vol. 1 CVPR, pp 511–518
Yang X, Jian L, Wu W, Liu K, Yan B, Zhou Z, Peng J (2019) Implementing real-time RCF-Retinex image enhancement method using CUDA. J Real-Time Image Process. https://doi.org/10.1007/s11554-018-0803-y
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Asan, M.A., Ozsoy, A. cuRCD: Region covariance descriptor CUDA implementation. Multimed Tools Appl 80, 19737–19751 (2021). https://doi.org/10.1007/s11042-021-10644-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11042-021-10644-2