任务感知双原型网络的人物交互少样本识别

doi:10.3778/j.issn.1673-9418.2008027

摘要/Abstract

摘要：

人物交互（HOI）识别是计算机视觉领域的重要研究热点。随着深度学习在图像分类任务中的巨大成功，人物交互识别任务也取得重大进展，但样本不平衡和组合爆炸问题仍是制约当前人物交互识别任务性能的关键挑战。由此，将人物交互识别任务与少样本学习相结合，将人物交互识别任务定义为一个少样本任务，并提出了任务感知双原型网络（TDP-Net）来解决少样本人物交互任务。具体地，首先使用图方法为每个任务生成语义感知的任务表示作为任务的先验信息，并使用语义图注意力模块（SGA-Module）生成注意力权重，对特征图中不同区域进行不同重要程度的关注，以适应不同任务条件下的映射关系，实现在新任务中自动推理。此外，还设计了一个双路原型模块（DP-Module）以分别产生交互类别的动作类原型和物体类原型，并分别对动词和名词进行分类。通过分别为动作和物体建立类原型，有效地分离了动作和物体间复杂的视觉关系。同时由于人物交互类别之间具有相似性，可通过重新组合动作和物体类别将知识迁移到新的交互类别中。实验结果表明，该模型在人物交互少样本任务上的平均准确率比基线方法在两个实验设置上分别提高了3.2个百分点和15.7个百分点，验证了TDP-Net在少样本人物交互任务中的有效性。

关键词: 计算机视觉, 图像分类, 人物交互（HOI）, 少样本学习（FSL）, 注意力机制

Abstract:

Recognizing human-object interaction (HOI) is an important research topic in computer vision. With the great success of deep learning in image classification, the HOI recognition task has also made great progress. However, the problems of instance imbalance and combinatorial explosion still remain the key challenges, which restrict the performance of HOI recognition methods. Therefore, this paper formulates HOI recognition in a few-shot scene to tackle the above problems and proposes a novel task-aware dual prototypical network (TDP-Net) to address few-shot HOI task. Specifically, it first assigns semantic-aware task representations for different tasks as their prior knowledge, subsequently generates attention weights by semantic graph attention module (SGA-Module). It effectively weights the importance on different regions of the visual features, adaptively for different task conditions, which realizes to reason for novel tasks. In addition, it designs a dual prototypes module (DP-Module) to generate both action class prototypes and object class prototypes, which classifies the verb and noun labels respectively. The complex visual relationships between actions and objects can be effectively separated by constructing class prototypes for actions and objects. Meanwhile, owing to the similarity among the related interactions, the knowledge is transferred to the new interactions by reorganizing the action and object prototypes. The experimental results show that the average accuracies of this model outperform the baseline by 3.2 percentage points and 15.7 percentage points on two exper-imental settings, which verifies its effectiveness on the few-shot HOI task.

Key words: computer vision, image classification, human-object interaction (HOI), few-shot learning (FSL), atten-tion mechanism

安平，冀中，刘西瑶. 任务感知双原型网络的人物交互少样本识别[J]. 计算机科学与探索, 2021, 15(11): 2184-2192.

AN Ping, JI Zhong, LIU Xiyao. Task-Aware Dual Prototypical Network for Few-Shot Human-Object Interaction Recognition[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(11): 2184-2192.

参考文献

[1] ZHU Y, ZHAO J K, WANG Y N, et al. A review of human action recognition based on deep learning[J]. Acta Automa-tica Sinica, 2016, 42(6): 848-857.
朱煜, 赵江坤, 王逸宁, 等. 基于深度学习的人体行为识别算法综述[J]. 自动化学报, 2016, 42(6): 848-857.
[2] GKIOXARI G, GIRSHICK R B, DOLLAR P, et al. Detecting and recognizing human-object interactions[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pat-tern Recognition, Salt Lake City, Jun 18-22, 2018. Washington: IEEE Computer Society, 2018: 8359-8367.
[3] QI S Y, WANG W G, JIA B X, et al. Learning human-object interactions by graph parsing neural networks[C]//LNCS 11213: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 407-423.
[4] VINYALS O, BLUNDELL C, LILLICRAP T P, et al. Match-ing networks for one shot learning[C]//Proceedings of the 29th Annual Conference on Neural Information Processing Systems, Barcelona, Dec 5-10, 2016. Red Hook: Curran Ass-ociates, 2016: 3630-3638.
[5] KOCH G, ZEMEL R, SALAKHUTDINOV R. Siamese ne-ural networks for one-shot image recognition[C]//Proceed-ings of the 32nd International Conference on Machine Lea-rning, Lille, Jul 6-11, 2015. Stroudsburg: ACL, 2015: 1-8.
[6] SUNG F, YANG Y X, ZHANG L, et al. Learning to compare: relation network for few-shot learning[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, Jun 18-22, 2018. Washington: IEEE Computer Society, 2018: 1199-1208.
[7] SNELL J, SWERSKY K, ZEMEL R S. Prototypical networks for few-shot learning[C]//Proceedings of the 2017 Annual Conference on Neural Information Processing Systems, Long Beach, Dec 4-9, 2017. Red Hook: Curran Associates, 2017: 4077-4087.
[8] FINN C, ABBEEL P, LEVINE S. Model-agnostic meta-learning for fast adaptation of deep networks[C]//Proceedings of the 34th International Conference on Machine Learning, Sydney, Aug 6-11, 2017: 1126-1135.
[9] NICHOL A, ACHIAM J, SCHULMAN J. On first-order meta-learning algorithms[J]. arXiv:1803.02999, 2018.
[10] WANG F, JIANG M Q, QIAN C, et al. Residual attention network for image classification[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 6450-6458.
[11] WU L, WANG Y, LI X, et al. Deep attention-based spatially recursive networks for fine-grained visual recognition[J]. IEEE Transactions on Systems, Man, and Cybernetics, 2019, 49(5): 1791-1802.
[12] WU S S, YANG J F, SHAN Y, et al. Research on generative adversarial networks using twins attention mechanism[J]. Journal of Frontiers of Computer Science and Technology, 2020, 14(5): 833-840.
武随烁, 杨金福, 单义, 等. 使用孪生注意力机制的生成对抗网络的研究[J]. 计算机科学与探索, 2020, 14(5): 833-840.
[13] FU J, LIU J, TIAN H J, et al. Dual attention network for scene segmentation[C]//Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, Jun 16-20, 2019. Piscataway: IEEE, 2018: 3146-3154.
[14] WEI Y C, FENG J S, LIANG X D, et al. Object region min-ing with adversarial erasing: a simple classification to sem-antic segmentation approach[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recogni-tion, Honolulu, Jul 21-26, 2017. Washington: IEEE Com-puter Society, 2017: 6488-6496.
[15] JI Z, XIONG K L, PANG Y W, et al. Video summarization with attention-based encoder-decoder networks[J]. IEEE Transactions on Circuits and Systems for Video Technology,2020, 30(6): 1709-1717.
[16] YU Y L, JI Z, FU Y W, et al. Stacked semantic-guided atten-tion model for fine-grained zero-shot learning[C]//Proceed-ings of the Annual Conference on Neural Information Pro-cessing Systems 2018, Montréal, Dec 3-8, 2018. Red Hook: Curran Associates, 2018: 5998-6007.
[17] XING C, ROSTAMZADEH N, ORESHKIN B N, et al. Ada-ptive cross-modal few-shot learning[J]. arXiv:1902.07104, 2019.
[18] JI Z, LI H H, HE Y Q. Zero-shot multi-label image classi-fication based on deep instance differentiation[J]. Journal of Frontiers of Computer Science and Technology, 2019, 13(1): 97-105.
冀中, 李慧慧, 何宇清. 基于深度示例差异化的零样本多标签图像分类[J]. 计算机科学与探索, 2019, 13(1): 97-105.
[19] JI Z, WANG H R, YU Y L, et al. A decadal survey of zero-shot image classification[J]. Science in China: Information Sciences, 2019, 49(10): 1299-1320.
冀中, 汪浩然, 于云龙, 等. 零样本图像分类综述: 十年进展[J]. 中国科学: 信息科学, 2019, 49(10): 1299-1320.
[20] CHAO Y W, WANG Z, HE Y G, et al. HICO: a benchmark for recognizing human-object interactions in images[C]// Proceedings of the 2015 IEEE International Conference on Computer Vision, Santiago, Dec 7-13, 2015. Washington:IEEE Computer Society, 2015: 1017-1025.
[21] LE D T, UIJLINGS J R R, BERNARDI R. TUHOI: trento universal human object interaction dataset[C]//Proceedings of the 3rd Workshop on Vision and Language, Dublin, Aug 23, 2014. Stroudsburg: ACL, 2014: 17-24.
[22] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Con-ference on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Soc-iety, 2016: 770-778.
[23] MIKOLOV T, SUTSKEVER I, CHEN K, et al. Distributed representations of words and phrases and their composition-ality[C]//Proceedings of the 27th Annual Conference on Neural Information Processing Systems 2013, Lake Tahoe, Dec 5-8, 2013. Red Hook: Curran Associates, 2013: 3111-3119.
[24] LI W B, WANG L, XU J L, et al. Revisiting local descriptor based image-to-class measure for few-shot learning[C]// Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, Jun 16-20, 2019. Piscataway: IEEE, 2019: 7260-7268.
[25] LIU L, ZHOU T Y, LONG G D, et al. Learning to propagate for graph meta-learning[C]//Proceedings of the 32nd Annual Conference on Neural Information Processing Systems, Van-couver, Dec 8-14, 2019. Red Hook: Curran Associates, 2019: 1037-1048.
[26] LI H Y, DONG W M, MEI X, et al. LGM-Net: learning to generate matching networks for few shot learning[C]//Pro-ceedings of the 36th International Conference on Machine Learning, Long Beach, Jun 10-15, 2019. New York: ACM, 2019: 3825-3834.