IoU Loss for 2D/3D Object Detection

By Dingfu Zhou, Jin Fang, Xibin Song, Chenye Guan, Junbo Yin, Yuchao Dai, Ruigang Yang. International Conference on 3D Vision, 2019

479 citations recorded by the source corpus.

In the 2D/3D object detection task, Intersection-over-Union (IoU) has been widely employed as an evaluation metric to evaluate the performance of different detectors in the testing stage. However, during the training stage, the common distance loss (e.g, L_1 or L_2) is often adopted as the loss function to minimize the discrepancy between the predicted and ground truth Bounding Box (Bbox). To eliminate the performance gap between training and testing, the IoU loss has been introduced for 2D object detection in [1] and [2]. Unfortunately, all these approaches only work for axis-aligned 2D Boxes, which cannot be applied for more general object detection task with rotated Boxes. To resolve this issue, we investigate the IoU computation for two rotated Boxes first and then implement a unified framework, IoU loss layer for both 2D and 3D object detection tasks. By integrating the implemented IoU loss into several state-of-the-art 3D object detectors, consistent improvements have been achieved for both bird-eye-view 2D detection and point cloud 3D detection on the public KITTI [3] benchmark.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/a7df9d524b89fb0e2133ef8ad9dab6f165d74c3a

How researchers see this paper

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

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Deep Learning for 3D Point Clouds: A Survey (2019, background): “[119] integrated the IoU of two 3D rotated bounding boxes into several stateof-the-art detectors [105], [109], [120] to achieve consistent performance improvement.”
Delving into Localization Errors for Monocular 3D Object Detection (2021, background): “Recently, some LiDAR based 3D detectors [40, 42] applied the IoU oriented optimization [31].”
SE-SSD: Self-Ensembling Single-Stage Object Detector From Point Cloud (2021, unknown): “49 470* 3D IoU Loss [37] 3DV 2019 LiDAR 86.”
TransFusion: Robust LiDAR-Camera Fusion for 3D Object Detection with Transformers (2022, background): “where Lcls is the binary cross entropy loss, Lreg is the L1 loss between the predicted BEV centers and the groundtruth centers (both normalized in [0, 1]), and Liou is the IoU loss [64] between the predicted boxes and ground-truth boxes.”
Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Clouds (2022, unknown): “4 3D IoU Loss [59] 3DV 2019 1-stage 86.”
PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection (2021, unknown): “28 3D IoU Loss [91] 3DV 2019 LiDAR only 86.”
YOLOv7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors (2022, methodology): “Being able to become a state-of-the-art real-time object detector usually requires the following characteristics: (1) a faster and stronger network architecture; (2) a more effective feature integration method [22, 97, 37, 74, 59, 30, 9, 45]; (3) a more accurate detection method [76, 77, 69]; (4) a ”
PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection (2019, unknown): “28 3D IoU Loss [39] 3DV 2019 LiDAR only 86.”
PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments (2020, methodology): “It should be noted that the proposed PIoU loss is different from [48] in which the IoU is computed based on axis alignment and polygon intersection, our method is more straightforward, i.”
YOLOv12: Attention-Centric Real-Time Object Detectors (2025, unknown): “Although improvements in YOLO have focused on areas such as loss functions [8, 35, 43, 44, 48, 67, 68], label assignment [22, 23, 34, 59, 69], network architecture design has remained a critical research priority [24, 28, 32, 57, 58].”
YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information (2024, background): “In addition, some researchers have tried to develop more general objective functions, such as loss function [5, 45, 46, 50, 77, 78], label assignment [10, 12, 33, 67, 79] and auxiliary supervision [18, 20, 24, 28, 29, 51, 54, 68, 76].”