{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2023,11,11]],"date-time":"2023-11-11T04:00:02Z","timestamp":1699675202815},"reference-count":32,"publisher":"Institute of Electronics, Information and Communications Engineers (IEICE)","issue":"24","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEICE Electron. Express"],"published-print":{"date-parts":[[2021,12,25]]},"DOI":"10.1587\/elex.18.20210491","type":"journal-article","created":{"date-parts":[[2021,12,1]],"date-time":"2021-12-01T22:07:53Z","timestamp":1638396473000},"page":"20210491-20210491","source":"Crossref","is-referenced-by-count":2,"title":["Dimension fusion: Dimension-level dynamically composable accelerator for convolutional neural networks"],"prefix":"10.1587","volume":"18","author":[{"given":"Huipeng","family":"Deng","sequence":"first","affiliation":[{"name":"School of Electronics and Information Technology, Sun Yat-sen University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jian","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Electronics and Information Technology, Sun Yat-sen University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Huafeng","family":"Ye","sequence":"additional","affiliation":[{"name":"School of Microelectronics Science and Technology, Sun Yat-sen University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shanlin","family":"Xiao","sequence":"additional","affiliation":[{"name":"School of Microelectronics Science and Technology, Sun Yat-sen University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xiangyu","family":"Meng","sequence":"additional","affiliation":[{"name":"School of Electronics and Information Technology, Sun Yat-sen University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhiyi","family":"Yu","sequence":"additional","affiliation":[{"name":"School of Microelectronics Science and Technology, Sun Yat-sen University"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"532","reference":[{"key":"1","doi-asserted-by":"crossref","unstructured":"[1] P. Zhou, <i>et al.<\/i>: \u201cState-clustering based multiple deep neural networks modeling approach for speech recognition,\u201d IEEE\/ACM Trans. Audio, Speech, Language Process. <b>23<\/b> (2015) 631 (DOI: 10.1109\/TASLP.2015.2392944).","DOI":"10.1109\/TASLP.2015.2392944"},{"key":"2","doi-asserted-by":"crossref","unstructured":"[2] A. DiSpirito, <i>et al.<\/i>: \u201cReconstructing undersampled photoacoustic microscopy images using deep learning,\u201d IEEE Trans. Med. Imag. <b>40<\/b> (2021) 562 (DOI: 10.1109\/TMI.2020.3031541).","DOI":"10.1109\/TMI.2020.3031541"},{"key":"3","doi-asserted-by":"crossref","unstructured":"[3] H. Shin, <i>et al.<\/i>: \u201cDeep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning,\u201d IEEE Trans. Med. Imag. <b>35<\/b> (2016) 1285 (DOI: 10.1109\/TMI.2016.2528162).","DOI":"10.1109\/TMI.2016.2528162"},{"key":"4","doi-asserted-by":"crossref","unstructured":"[4] A.H. Fan, <i>et al.<\/i>: \u201cF-C3D: FPGA-based 3-dimensional convolutional neural network,\u201d 2017 27th International Conference on Field Programmable Logic and Applications (2017) 1 (DOI: 10.23919\/FPL.2017.8056779).","DOI":"10.23919\/FPL.2017.8056779"},{"key":"5","doi-asserted-by":"crossref","unstructured":"[5] Y. Wang, <i>et al.<\/i>: \u201cAn edge 3D CNN accelerator for low-power activity recognition,\u201d IEEE Trans. Comput.-Aided Design Integr. Circuits Syst. <b>40<\/b> (2021) 918 (DOI: 10.1109\/TCAD.2020.3011042).","DOI":"10.1109\/TCAD.2020.3011042"},{"key":"6","doi-asserted-by":"crossref","unstructured":"[6] K. Hegde, <i>et al.<\/i>: \u201cMorph: flexible acceleration for 3D CNN-based video understanding,\u201d 2018 51st Annual IEEE\/ACM International Symposium on Microarchitecture (MICRO) (2018) 933 (DOI: 10.1109\/MICRO.2018.00080).","DOI":"10.1109\/MICRO.2018.00080"},{"key":"7","doi-asserted-by":"crossref","unstructured":"[7] H. Wu, <i>et al.<\/i>: \u201cSpatiotemporal multimodal learning with 3D CNNs for video action recognition,\u201d IEEE Trans. Circuits Syst. Video Technol. (2021) 1 (DOI: 10.1109\/TCSVT.2021.3077512).","DOI":"10.1109\/TCSVT.2021.3077512"},{"key":"8","doi-asserted-by":"crossref","unstructured":"[8] K. Ranipa, <i>et al.<\/i>: \u201cMultimodal CNN fusion architecture with multi-features for heart sound classification,\u201d 2021 IEEE International Symposium on Circuits and Systems (2021) 1 (DOI: 10.1109\/ISCAS51556.2021.9401551).","DOI":"10.1109\/ISCAS51556.2021.9401551"},{"key":"9","doi-asserted-by":"crossref","unstructured":"[9] A. Torfi, <i>et al.<\/i>: \u201c3D convolutional neural networks for cross audio-visual matching recognition,\u201d IEEE Access <b>5<\/b> (2017) 22081 (DOI: 10.1109\/ACCESS.2017.2761539).","DOI":"10.1109\/ACCESS.2017.2761539"},{"key":"10","unstructured":"[10] S. Ioffe and C. Szegedy: \u201cBatch normalization: accelerating deep network training by reducing internal covariate shift,\u201d Proceedings of the 32nd International Conference on International Conference on Machine Learning (2015) 448."},{"key":"11","doi-asserted-by":"crossref","unstructured":"[11] D. Tran, <i>et al.<\/i>: \u201cA closer look at spatiotemporal convolutions for action recognition,\u201d 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (2018) 6450 (DOI: 10.1109\/CVPR.2018.00675).","DOI":"10.1109\/CVPR.2018.00675"},{"key":"12","doi-asserted-by":"crossref","unstructured":"[12] W. Xu, <i>et al.<\/i>: \u201cReconfigurable and low-complexity accelerator for convolutional and generative networks over finite fields,\u201d IEEE Trans. Comput.-Aided Design Integr. Circuits Syst. <b>39<\/b> (2020) 4894 (DOI: 10.1109\/TCAD.2020.2973355).","DOI":"10.1109\/TCAD.2020.2973355"},{"key":"13","doi-asserted-by":"crossref","unstructured":"[13] S. Kala, <i>et al.<\/i>: \u201cHigh-performance CNN accelerator on FPGA using unified winograd-GEMM architecture,\u201d IEEE Trans. Very Large Scale Integr. (VLSI) Syst. <b>27<\/b> (2019) 2816 (DOI: 10.1109\/TVLSI.2019.2941250).","DOI":"10.1109\/TVLSI.2019.2941250"},{"key":"14","doi-asserted-by":"crossref","unstructured":"[14] H. Ye, <i>et al.<\/i>: \u201cHybridDNN: a framework for high-performance hybrid DNN accelerator design and implementation,\u201d 2020 57th ACM\/IEEE Design Automation Conference (DAC) (2020) 1 (DOI: 10.1109\/DAC18072.2020.9218684).","DOI":"10.1109\/DAC18072.2020.9218684"},{"key":"15","doi-asserted-by":"crossref","unstructured":"[15] C. Yang, <i>et al.<\/i>: \u201cWRA: a 2.2-to-6.3 TOPS highly unified dynamically reconfigurable accelerator using a novel winograd decomposition algorithm for convolutional neural networks,\u201d IEEE Trans. Circuits Syst. I, Reg. Papers <b>66<\/b> (2019) 3480 (DOI: 10.1109\/TCSI.2019.2928682).","DOI":"10.1109\/TCSI.2019.2928682"},{"key":"16","doi-asserted-by":"crossref","unstructured":"[16] C. Yang, <i>et al.<\/i>: \u201cA stride-based convolution decomposition method to stretch CNN acceleration algorithms for efficient and flexible hardware implementation,\u201d IEEE Trans. Circuits Syst. I, Reg. Papers <b>67<\/b> (2020) 3007 (DOI: 10.1109\/TCSI.2020.2985727).","DOI":"10.1109\/TCSI.2020.2985727"},{"key":"17","doi-asserted-by":"crossref","unstructured":"[17] J. Pan and D. Chen: \u201cAccelerate non-unit stride convolutions with Winograd algorithms,\u201d 2021 26th Asia and South Pacific Design Automation Conference (2021) 358 (DOI: 10.1145\/3394885.3431534).","DOI":"10.1145\/3394885.3431534"},{"key":"18","doi-asserted-by":"crossref","unstructured":"[18] J. Shen, <i>et al.<\/i>: \u201cToward an efficient deep pipelined template-based architecture for accelerating the entire 2-D and 3-D CNNs on FPGA,\u201d IEEE Trans. Comput.-Aided Design Integr. Circuits Syst. <b>39<\/b> (2020) 1442 (DOI: 10.1109\/TCAD.2019.2912894).","DOI":"10.1109\/TCAD.2019.2912894"},{"key":"19","unstructured":"[19] J. Jiang, <i>et al.<\/i>: \u201cA reconfigurable Winograd CNN accelerator with nesting decomposition algorithm for computing convolution with large filters,\u201d arXiv preprint (2021) arXiv: 2102.13272."},{"key":"20","doi-asserted-by":"crossref","unstructured":"[20] H. Deng, <i>et al.<\/i>: \u201c3D-VNPU: a flexible accelerator for 2D\/3D CNNs on FPGA,\u201d 2021 IEEE 29th Annual International Symposium on Field-Programmable Custom Computing Machines (2021) 181 (DOI: 10.1109\/FCCM51124.2021.00029).","DOI":"10.1109\/FCCM51124.2021.00029"},{"key":"21","doi-asserted-by":"crossref","unstructured":"[21] A. Lavin and S. Gray: \u201cFast algorithms for convolutional neural networks,\u201d 2016 IEEE Conference on Computer Vision and Pattern Recognition (2016) 4013 (DOI: 10.1109\/CVPR.2016.435).","DOI":"10.1109\/CVPR.2016.435"},{"key":"22","doi-asserted-by":"crossref","unstructured":"[22] J. Yepez and S. Ko: \u201cStride 2 1-D, 2-D, and 3-D Winograd for convolutional neural networks,\u201d IEEE Trans. Very Large Scale Integr. (VLSI) Syst. <b>28<\/b> (2020) 853 (DOI: 10.1109\/TVLSI.2019.2961602).","DOI":"10.1109\/TVLSI.2019.2961602"},{"key":"23","unstructured":"[23] J. Wang, <i>et al.<\/i>: \u201cFWUA: a flexible winograd-based uniform accelerator for 1D\/2D\/3D CNNs,\u201d submitted to The 4th IEEE International Conference on Integrated Circuits, Technologies and Applications."},{"key":"24","doi-asserted-by":"crossref","unstructured":"[24] K. He, <i>et al.<\/i>: \u201cDeep residual learning for image recognition,\u201d 2016 IEEE Conference on Computer Vision and Pattern Recognition (2016) 770 (DOI: 10.1109\/CVPR.2016.90).","DOI":"10.1109\/CVPR.2016.90"},{"key":"25","unstructured":"[25] A. Krizhevsky, <i>et al.<\/i>: \u201cImageNet classification with deep convolutional neural networks,\u201d Proceedings of the 25th International Conference on Neural Information Processing Systems (2012) 1097"},{"key":"26","doi-asserted-by":"crossref","unstructured":"[26] K. Hara, <i>et al.<\/i>: \u201cCan spatiotemporal 3D CNNs retrace the history of 2D CNNs and ImageNet?,\u201d 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (2018) 6546 (DOI: 10.1109\/CVPR.2018.00685).","DOI":"10.1109\/CVPR.2018.00685"},{"key":"27","doi-asserted-by":"crossref","unstructured":"[27] J. Carreira and A. Zisserman: \u201cQuo vadis, action recognition? A new model and the kinetics dataset,\u201d 2017 IEEE Conference on Computer Vision and Pattern Recognition (2017) 4724 (DOI: 10.1109\/CVPR.2017.502).","DOI":"10.1109\/CVPR.2017.502"},{"key":"28","doi-asserted-by":"crossref","unstructured":"[28] M. Sun, <i>et al.<\/i>: \u201c3D CNN acceleration on FPGA using hardware-aware pruning,\u201d 2020 57th ACM\/IEEE Design Automation Conference (DAC) (2020) 1 (DOI: 10.1109\/DAC18072.2020.9218571).","DOI":"10.1109\/DAC18072.2020.9218571"},{"key":"29","unstructured":"[29] K. Soomro, <i>et al.<\/i>: \u201cUCF101: a dataset of 101 human action classes from videos in the wild,\u201d CRCV-TR-12-01 (2012)."},{"key":"30","doi-asserted-by":"crossref","unstructured":"[30] O. Russakovsky, <i>et al.<\/i>: \u201cImageNet large scale visual recognition challenge,\u201d Int. J. Comput. Vis. <b>115<\/b> (2015) 211 (DOI: 10.1007\/s11263-015-0816-y).","DOI":"10.1007\/s11263-015-0816-y"},{"key":"31","doi-asserted-by":"crossref","unstructured":"[31] J. Lee, <i>et al.<\/i>: \u201cUNPU: an energy-efficient deep neural network accelerator with fully variable weight bit precision,\u201d IEEE J. Solid-State Circuits <b>54<\/b> (2019) 173 (DOI: 10.1109\/JSSC.2018.2865489).","DOI":"10.1109\/JSSC.2018.2865489"},{"key":"32","doi-asserted-by":"crossref","unstructured":"[32] L. Rao, <i>et al.<\/i>: \u201cAn energy-efficient accelerator for rain removal based on convolutional neural network,\u201d IEEE Trans. Circuits Syst. II, Exp. Briefs <b>68<\/b> (2021) 2957 (DOI: 10.1109\/TCSII.2021.3071455).","DOI":"10.1109\/TCSII.2021.3071455"}],"container-title":["IEICE Electronics Express"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.jstage.jst.go.jp\/article\/elex\/18\/24\/18_18.20210491\/_pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,12,25]],"date-time":"2021-12-25T05:12:57Z","timestamp":1640409177000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.jstage.jst.go.jp\/article\/elex\/18\/24\/18_18.20210491\/_article"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,12,25]]},"references-count":32,"journal-issue":{"issue":"24","published-print":{"date-parts":[[2021]]}},"URL":"https:\/\/doi.org\/10.1587\/elex.18.20210491","relation":{},"ISSN":["1349-2543"],"issn-type":[{"value":"1349-2543","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,12,25]]},"article-number":"18.20210491"}}