Laser & Optoelectronics Progress, Volume. 56, Issue 22, 221003(2019)
Light-Weight Multi-Object Detection Network Based on Inverted Residual Structure
Figures & Tables(13)
Fig. 1. Decoupling process of the depth separable convolution. (a) Standard convolution; (b) depth separable convolution
Fig. 2. Residual block and inverted residual block. (a) Residual block; (b) inverted residual block when stride is 1
Fig. 6. Comparison of detection results. (a)(d) Original input images ; (b)(e) detection results with YOLOv3-Tiny Model; (c)(f) detection results with IR-YOLO Model
Table 1. Parameters of inverted residual block
View tableTable 1. Parameters of inverted residual block
Input Operation Output h ×w ×k 1×1 point conv, ReLU h ×w ×2k h ×w ×2k 3×3/ s depth conv, ReLU × ×2 k × ×2 k 1×1 point conv, linear × ×2 k
Table 2. VOC dataset
View tableTable 2. VOC dataset
Category Train set Test set Aeroplane 1171 285 Bicycle 1064 337 Bird 1605 459 Boat 1140 263 Bottle 1764 469 Bus 822 213 Car 3267 1201 Cat 1593 358 Chair 3152 756 Cow 847 244 Dining table 824 206 Dog 2025 489 Horse 1072 348 Motor bike 1052 325 Person 13256 4528 Potted plant 1487 480 Sheep 1070 242 Sofa 814 239 Train 925 282 TV monitor 1108 308 Total 40058 12032
Table 3. Hyper parameters
View tableTable 3. Hyper parameters
Parameters name Value Batch 64 Momentum 0.9 Weight decay 0.0005 Learning rate 0.001
Table 4. Comparison on number of floating point operations
View tableTable 4. Comparison on number of floating point operations
Input Output Number of floating point operations in standard conv /109 Number of floating point operations in inverted residual block /109 Expand point conv Depth conv Squeeze point conv 208×208×16 208×208×32 0.399 0.044 0.025 0.089 104×104×32 104×104×64 0.399 0.044 0.012 0.089 52×52×64 52×52×128 0.399 0.044 0.006 0.089 26×26×128 26×26×256 0.399 0.044 0.003 0.089 13×13×256 13×13×512 0.399 0.044 0.002 0.089 13×13×512 13×13×1024 1.595 0.177 0.003 0.354
Table 5. Comparison detection speed of IR-YOLO model and YOLOv3-Tiny model
View tableTable 5. Comparison detection speed of IR-YOLO model and YOLOv3-Tiny model
Model CPU speed / (frame·s-1) GPU speed / (frame·s-1) YOLOv3-Tiny 1.2 31.3 IR-YOLO 1.7 31.2
Table 6. Comparison mAP of different training numbers
View tableTable 6. Comparison mAP of different training numbers
Training number YOLOv3-Tiny mAP /% IR-YOLO mAP /% 65000 45.15 43.33 75000 45.60 44.37 85000 45.17 45.23 90000 42.75 44.20 95000 42.76 46.07
Table 7. Comparison of detection results of IR-YOLO and YOLOv3-Tiny on VOC dataset%
View tableTable 7. Comparison of detection results of IR-YOLO and YOLOv3-Tiny on VOC dataset%
Category YOLOv3-Tiny IR-YOLO Aeroplane 54.78 56.38 Bicycle 60.79 57.86 Bird 27.24 28.19 Boat 27.9 28.92 Bottle 14.8 17.58 Bus 56.98 58.48 Car 63.8 64.05 Cat 50.39 53.57 Chair 25.77 23.25 Cow 46.43 45.48 Dining table 39.66 45.48 Dog 46.09 45.68 Horse 66.62 62.45 Motor bike 64.09 62.85 Person 59.23 59.4 Potted plant 18.22 17.22 Sheep 47.57 44.68 Sofa 39.39 43.11 Train 54.02 58.25 TV monitor 50.34 48.62 mAP 45.60 46.07
Tools
Get Citation
Copy Citation Text
Wanjun Liu, Mingyue Gao, Haicheng Qu, Lamei Liu. Light-Weight Multi-Object Detection Network Based on Inverted Residual Structure[J]. Laser & Optoelectronics Progress, 2019, 56(22): 221003
Paper Information
Category: Image Processing
Received: Mar. 27, 2019
Accepted: May. 17, 2019
Published Online: Nov. 2, 2019
The Author Email: Qu Haicheng (quhaicheng@lntu.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20