A compression pipeline for one-stage object detection model

Deep neural networks (DNNs) have strong fitting ability on a variety of computer vision tasks, but they also require intensive computing power and large storage space, which are not always available in portable smart devices. Although a lot of studies have contributed to the compression of image cla...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of real-time image processing Ročník 18; číslo 6; s. 1949 - 1962
Hlavní autoři: Li, Zhishan, Sun, Yiran, Tian, Guanzhong, Xie, Lei, Liu, Yong, Su, Hongye, He, Yifan
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.12.2021
Springer Nature B.V
Témata:
Accuracy
Algorithms
Artificial neural networks
Classification
Computer Graphics
Computer Science
Computer vision
Distillation
Floating point arithmetic
Image classification
Image compression
Image Processing and Computer Vision
Integers
Knowledge
Knowledge management
Learning
Methods
Multimedia Information Systems
Neural networks
Object recognition
Original Research Paper
Pattern Recognition
Portable equipment
Signal,Image and Speech Processing
Training
Model compression pipeline
Real-time
Object detection
ISSN:1861-8200, 1861-8219
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Deep neural networks (DNNs) have strong fitting ability on a variety of computer vision tasks, but they also require intensive computing power and large storage space, which are not always available in portable smart devices. Although a lot of studies have contributed to the compression of image classification networks, there are few model compression algorithms for object detection models. In this paper, we propose a general compression pipeline for one-stage object detection networks to meet the real-time requirements. Firstly, we propose a softer pruning strategy on the backbone to reduce the number of filters. Compared with original direct pruning, our method can maintain the integrity of network structure and reduce the drop of accuracy. Secondly, we transfer the knowledge of the original model to the small model by knowledge distillation to reduce the accuracy drop caused by pruning. Finally, as edge devices are more suitable for integer operations, we further transform the 32-bit floating point model into the 8-bit integer model through quantization. With this pipeline, the model size and inference time are compressed to 10% or less of the original, while the mAP is only reduced by 2.5% or less. We verified that performance of the compression pipeline on the Pascal VOC dataset.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1861-8200
1861-8219
DOI:10.1007/s11554-020-01053-z