A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection

被引:1097
作者
Cai, Zhaowei [1 ]
Fan, Quanfu [2 ]
Feris, Rogerio S. [2 ]
Vasconcelos, Nuno [1 ]
机构
[1] Univ Calif San Diego, SVCL, San Diego, CA 92103 USA
[2] IBM TJ Watson Res, Yorktown Hts, NY USA
来源
COMPUTER VISION - ECCV 2016, PT IV | 2016年 / 9908卷
基金
美国国家科学基金会;
关键词
Object detection; Multi-scale; Unified neural network;
D O I
10.1007/978-3-319-46493-0_22
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
A unified deep neural network, denoted the multi-scale CNN (MS-CNN), is proposed for fast multi-scale object detection. The MS-CNN consists of a proposal sub-network and a detection sub-network. In the proposal sub-network, detection is performed at multiple output layers, so that receptive fields match objects of different scales. These complementary scale-specific detectors are combined to produce a strong multi-scale object detector. The unified network is learned end-to-end, by optimizing a multi-task loss. Feature upsampling by deconvolution is also explored, as an alternative to input upsampling, to reduce the memory and computation costs. State-of-the-art object detection performance, at up to 15 fps, is reported on datasets, such as KITTI and Caltech, containing a substantial number of small objects.
引用
收藏
页码:354 / 370
页数:17
相关论文
共 50 条
[31]   A Novel Multi-Scale Feature Fusion Method for Region Proposal Network in Fast Object Detection [J].
Liu, Gang ;
Wang, Chuyi .
INTERNATIONAL JOURNAL OF DATA WAREHOUSING AND MINING, 2020, 16 (03) :132-145
[32]   AFSNet: Multi-Scale Adaptive Feature Scaling Convolutional Network for Real-time Object Detection [J].
Hague, Md Foysal ;
Kang, Dae-Seong .
INTERNATIONAL JOURNAL OF COMPUTER SCIENCE AND NETWORK SECURITY, 2020, 20 (06) :216-222
[33]   Selective Multi-scale Learning for Object Detection [J].
Chen, Junliang ;
Lu, Weizeng ;
Shen, Linlin .
ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING - ICANN 2021, PT II, 2021, 12892 :3-14
[34]   A 1.15-TOPS 6.57-TOPS/W Neural Network Processor for Multi-Scale Object Detection With Reduced Convolutional Operations [J].
Kawamoto, Reiya ;
Taichi, Masakazu ;
Kabuto, Masaya ;
Watanabe, Daisuke ;
Izumi, Shintaro ;
Yoshimoto, Masahiko ;
Kawaguchi, Hiroshi ;
Matsukawa, Go ;
Goto, Toshio ;
Kojima, Motoshi .
IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING, 2020, 14 (04) :634-645
[35]   Multi-Scale Feature Selective Matching Network for Object Detection [J].
Pei, Yuanhua ;
Dong, Yongsheng ;
Zheng, Lintao ;
Ma, Jinwen .
MATHEMATICS, 2023, 11 (12)
[36]   Image Classification Method Based on Multi-Scale Convolutional Neural Network [J].
Du, Shaobo ;
Li, Jing .
JOURNAL OF CIRCUITS SYSTEMS AND COMPUTERS, 2024, 33 (10)
[37]   Multi-Scale Acoustic Velocity Inversion Based on a Convolutional Neural Network [J].
Li, Wenda ;
Wu, Tianqi ;
Liu, Hong .
REMOTE SENSING, 2024, 16 (05)
[38]   Human and object detection using Hybrid Deep Convolutional Neural Network [J].
Mukilan, P. ;
Semunigus, Wogderess .
SIGNAL IMAGE AND VIDEO PROCESSING, 2022, 16 (07) :1913-1923
[39]   Human and object detection using Hybrid Deep Convolutional Neural Network [J].
P. Mukilan ;
Wogderess Semunigus .
Signal, Image and Video Processing, 2022, 16 :1913-1923
[40]   Chicken Image Segmentation via Multi-Scale Attention-Based Deep Convolutional Neural Network [J].
Li, Wei ;
Xiao, Yang ;
Song, Xibin ;
Lv, Na ;
Jiang, Xinbo ;
Huang, Yan ;
Peng, Jingliang .
IEEE ACCESS, 2021, 9 :61398-61407