Generalized Focal Loss V2: Learning Reliable Localization Quality Estimation for Dense Object Detection

By Xiang-Yang Li, Wenhai Wang, Xiaolin Hu, Jun Yu Li, Jinhui Tang, Jian Yang. 2020

334 citations recorded by the source corpus.

Localization Quality Estimation (LQE) is crucial and popular in the recent advancement of dense object detectors since it can provide accurate ranking scores that benefit the Non-Maximum Suppression processing and improve detection performance. As a common practice, most existing methods predict LQE scores through vanilla convolutional features shared with object classification or bounding box regression. In this paper, we explore a completely novel and different perspective to perform LQE – based on the learned distributions of the four parameters of the bounding box. The bounding box distributions are inspired and introduced as "General Distribution" in GFLV1, which describes the uncertainty of the predicted bounding boxes well. Such a property makes the distribution statistics of a bounding box highly correlated to its real localization quality. Specifically, a bounding box distribution with a sharp peak usually corresponds to high localization quality, and vice versa. By leveraging the close correlation between distribution statistics and the real localization quality, we develop a considerably lightweight Distribution-Guided Quality Predictor (DGQP) for reliable LQE based on GFLV1, thus producing GFLV2. To our best knowledge, it is the first attempt in object detection to use a highly relevant, statistical representation to facilitate LQE. Extensive experiments demonstrate the effectiveness of our method. Notably, GFLV2 (ResNet101) achieves 46.2 AP at 14.6 FPS, surpassing the previous state-of-the-art ATSS baseline (43.6 AP at 14.6 FPS) by absolute 2.6 AP on COCO test-dev, without sacrificing the efficiency both in training and inference.

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

Overcoming Catastrophic Forgetting in Incremental Object Detection via Elastic Response Distillation (2022, background): “In contrast, currently, the classic training paradigm of object detection models [19,33] does not have such capability.”
YOLOv6: A Single-Stage Object Detection Framework for Industrial Applications (2022, background): “Upon DFL, DFLv2 [19] develops a lightweight sub-network to leverage the close correlation between distribution statistics and the real localization quality, which further boosts the detection performance.”
YOLOv10: Real-Time End-to-End Object Detection (2024, background): “Over the past years, substantial efforts have been directed towards developing efficient detectors [19, 57, 48, 34, 79, 75, 32, 31, 41].”
TOOD: Task-aligned One-stage Object Detection (2021, methodology): “We use different values of m in [5, 9, 13, 17, 21], and achieve results in a range of 42.”
Simple Multi-dataset Detection (2021, background): “This fragmentation brought rapid progress in object detection [5, 10, 20, 31, 39, 45] and instance segmentation [14], but comes with a drawback.”
Localization Distillation for Object Detection (2022, methodology): “work [37], the LD models with the 1333× [480 : 960] multiscale2× training schedule (24 epochs) are included.”
Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions (2021, background): “However, when the input scale does not exceed 640 × 640 pixels, the GFLOPs of PVT-Small and ResNet50 are similar.”
Hyper-YOLO: When Visual Object Detection Meets Hypergraph Computation (2024, neutral): “The single-stage design of YOLO simpliﬁes the object detection pipeline, enabling high performance with relatively small model sizes, which makes YOLO [23], [24] a popular choice for practical, real-time scenarios where efﬁciency and scalability are paramount.”
YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors (2022, methodology): “However, in recent years, if we take object detection as an example, researchers often use the quality and distribution of prediction output by the network, and then consider together with the ground truth to use some calculation and optimization methods to generate a reliable soft label [61, 8, 36,”
Probabilistic two-stage detection (2021, methodology): “While one-stage detectors have achieved competitive performance (Zhang et al., 2020b; Kim & Lee, 2020; Zhang et al., 2019; Li et al., 2020b; Zhu et al., 2020a), they usually rely on heavier separate classiﬁcation and regression branches than two-stage models.”