Learning Rich Features at High-Speed for Single-Shot Object Detection

By Tiancai Wang, Rao Muhammad Anwer, Hisham Cholakkal, Fahad Shahbaz Khan, Yanwei Pang, Ling Shao. 2019

97 citations recorded by the source corpus.

Single-stage object detection methods have received significant attention recently due to their characteristic realtime capabilities and high detection accuracies. Generally, most existing single-stage detectors follow two common practices: they employ a network backbone that is pretrained on ImageNet for the classification task and use a top-down feature pyramid representation for handling scale variations. Contrary to common pre-training strategy, recent works have demonstrated the benefits of training from scratch to reduce the task gap between classification and localization, especially at high overlap thresholds. However, detection models trained from scratch require significantly longer training time compared to their typical finetuning based counterparts. We introduce a single-stage detection framework that combines the advantages of both fine-tuning pretrained models and training from scratch. Our framework constitutes a standard network that uses a pre-trained backbone and a parallel light-weight auxiliary network trained from scratch. Further, we argue that the commonly used top-down pyramid representation only focuses on passing high-level semantics from the top layers to bottom layers. We introduce a bi-directional network that efficiently circulates both low-/mid-level and high-level semantic information in the detection framework. Experiments are performed on MS COCO and UAVDT datasets. Compared to the baseline, our detector achieives an absolute gain of 7.4% and 4.2% in average precision (AP) on MS COCO and UAVDT datasets, respectively using VGG backbone. For a 300×300 input on the MS COCO test set, our detector with ResNet backbone surpasses existing single-stage detection methods for single-scale inference achieving 34.3 AP, while operating at an inference time of 19 milliseconds on a single Titan X GPU. Code is avail- able at https://github.com/vaesl/LRF-Net.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

Bridging the Gap Between Anchor-Based and Anchor-Free Detection via Adaptive Training Sample Selection (2019, background): “Thereafter, plenty of works are presented to boost its performance in different aspects, such as fusing context information from different layers [24, 12, 69], training from scratch [50, 73], introducing new loss function [33, 6], anchor refinement and matching [66, 67], architecture redesign [21, 2”
Incremental Object Detection via Meta-Learning (2020, neutral): “Extending our methodology to single stage detectors [57], [58], [59] and incremental versions of related problem settings such as action recognition [60], object counting [61] and instance segmentation [62] are interesting and important research directions.”
MiniVLM: A Smaller and Faster Vision-Language Model (2020, background): “Region feature is built on top of the object detector, and the detector can be two-stage [38, 8, 24, 55] or one-stage [37, 48, 29, 27, 6, 58, 53, 49].”
YOLOv4: Optimal Speed and Accuracy of Object Detection (2020, background): “Learning Rich Features at High-Speed for Single-Shot Object Detection [84] LRF VGG-16 300 76.”
SipMask: Spatial Information Preservation for Fast Image and Video Instance Segmentation (2020, background): “Deep learning has achieved great success in a variety of computer vision tasks [20,12,43,35,45,44,25,24,53,52].”
D2Det: Towards High Quality Object Detection and Instance Segmentation (2020, unknown): “RetinaNet [32] LRF-Net [47] FPN [31] NDFT [48] D2Det”
CSPNet: A New Backbone that can Enhance Learning Capability of CNN (2019, methodology): “If compared to state-of-the-art LRF [38] under the input image size 512 × 512, CSPResNeXt50 with PANet (SPP) outperforms ResNet101 with LRF by 0.7% AP, 1.5% AP 50 and 1.1% AP 75 .”