Object Detection in RGB-D Image Based on ANNet

Abstract

Abstract: The wide spread of depth images acquisition devices makes object detection in RGB-D images a hotspot in the field of computer vision. In order to make the features extracted by CNN more robust and to improve the detection accuracy, an improved CNN called ANNet was designed. To enhance the model discriminability of local patches within the receptive field, some linear convolutional layers in the AlexNet with nonlinear convolutional layers were replaced which contained multilayer perceptron against the linear feature between convolution filter and underlying data patch. The experiment result shows that the detection accuracy is improved by 3% in the RGB images and 4% in the RGB-D images on the NYUD2 datasets using the improved network.

Key words: object detection, convolutional neural network, AlexNet, RGB-D images

CLC Number:

TP391

Cai Qiang, Wei Liwei, Li Haisheng, Cao Jian. Object Detection in RGB-D Image Based on ANNet[J]. Journal of System Simulation, 2016, 28(9): 2260-2266.

References 18

[1]	黄凯奇, 任伟强, 谭铁牛. 图像物体分类与检测算法综述[J]. 计算机学报, 2014, 37(6): 1225-1240.
[2]	Belongie S, Malik J, Puzicha J.Shape matching and object recognition using shape contexts[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence (S0162-8828), 2002, 24(4): 509-522.
[3]	Leung T, Malik J.Representing and recognizing the visual appearance of materials using three-dimensional textons[J]. International Journal of Computer Vision (S0920-5691), 2001, 43(1): 29-44.
[4]	Bo L, Ren X, Fox D.Depth kernel descriptors for object recognition[C]// Proceedings of Intelligent Robots and Systems (IROS), 2011 IEEE. France: IEEE, 2011: 821-826.
[5]	Cheng Y, Cai R, Zhao X, et al.Convolutional fisher kernels for rgb-d object recognition[C]// Proceedings of 2015 International Conference on 3D Vision (3DV). France: IEEE, 2015: 135-143.
[6]	Song S, Xiao J. Deep Sliding Shapes for Amodal 3D Object Detection in RGB-D Images [J]. arXiv preprint arXiv:1511.02300, 2015.
[7]	Gupta S, Girshick R, Arbeláez P, et al.Learning rich features from RGB-D images for object detection and segmentation[M]// Computer Vision-ECCV 2014. Zurich, Switzerland: Springer, 2014: 345-360.
[8]	Girshick R, Donahue J, Darrell T, et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Columbus, USA: IEEE, 2014: 580-587.
[9]	Girshick R.Fast r-cnn[C]// Proceedings of the IEEE International Conference on Computer Vision. 2015: 1440-1448, Santiago, Chile. USA: IEEE, 2015: 91-99.
[10]	Ren S, He K, Girshick R, et al.Faster R-CNN: Towards real-time object detection with region proposal networks[C]// Advances in Neural Information Processing Systems. Quebec, Canada: NIPS, 2015: 91-99.
[11]	LeCun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE (S0018-9219), 1998, 86(11): 2278-2324.
[12]	Krizhevsky A, Sutskever I, Hinton G E.Imagenet classification with deep convolutional neural networks[C]// Advances in Neural Information Processing Systems. Navada, USA: NIPS, 2012: 1097-1105.
[13]	Zeiler M D, Fergus R.Visualizing and understanding convolutional networks [M]// Computer vision-ECCV 2014. Germany: Springer International Publishing, 2014: 818-833.
[14]	Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition [J]. arXiv preprint arXiv:1409.1556, 2014.
[15]	Szegedy C, Liu W, Jia Y, et al.Going deeper with convolutions[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston, USA: IEEE, 2015: 1-9
[16]	Ouyang W, Luo P, Zeng X, et al. Deepid-net: multi-stage and deformable deep convolutional neural networks for object detection [J]. arXiv preprint arXiv:1409.3505, 2014
[17]	Lin M, Chen Q, Yan S. Network in network [J]. arXiv preprint arXiv:1312.4400, 2013.
[18]	Silberman N, Hoiem D, Kohli P, et al.Indoor Segmentation and Support Inference from RGB-D Images[M]// Computer Vision-ECCV 2012. Berlin, Heidelberg, Germany: Springer, 2012: 746-76.