EfficientDet: Scalable and Efficient Object Detection
Authors:
- Mingxing TanMTMingxing Tan
- Quoc V. LeQLQuoc V. Le
Preprint
Abstract
Model efficiency has become increasingly important in computer vision. In this paper, we systematically study neural network architecture design choices for object detection and propose several key optimizations to improve efficiency. First, we propose a weighted bi-directional feature pyramid network (BiFPN), which allows easy and fast multiscale feature fusion; Second, we propose a compound scaling method that uniformly scales the resolution, depth, and width for all backbone, feature network, and box/class prediction networks at the same time. Based on these optimizations and better backbones, we have developed a new family of object detectors, called EfficientDet, which consistently achieve much better efficiency than prior art across a wide spectrum of resource constraints. In particular, with single model and single-scale, our EfficientDet-D7 achieves state-of-the-art 55.1 AP on COCO test-dev with 77M parameters and 410B FLOPs, being 4x - 9x smaller and using 13x - 42x fewer FLOPs than previous detectors. Code is available at https://github.com/google/automl/tree/master/efficientdet.Keywords
All Related Versions
This publication has 21 references indexed in Scilit:
- Learning Data Augmentation Strategies for Object Detection2020
- YOLO-LITE: A Real-Time Object Detection Algorithm Optimized for Non-GPU Computers2018
- Sigmoid-weighted linear units for neural network function approximation in reinforcement learningNeural Networks, 2018
- CornerNet: Detecting Objects as Paired Keypoints2018
- Deep Feature Pyramid Reconfiguration for Object Detection2018
- Parallel Feature Pyramid Network for Object Detection2018
- Feature Pyramid Networks for Object Detection2017
- Speed/Accuracy Trade-Offs for Modern Convolutional Object Detectors2017
- Aggregated Residual Transformations for Deep Neural Networks2017
- Deep Residual Learning for Image Recognition2016