UnitBox: An Advanced Object Detection Network

By Jiahui Yu, Yuning Jiang, Zhangyang Wang, Zhimin Cao, Thomas S. Huang. 2016

1,757 citations recorded by the source corpus.

In present object detection systems, the deep convolutional neural networks (CNNs) are utilized to predict bounding boxes of object candidates, and have gained performance advantages over the traditional region proposal methods. However, existing deep CNN methods assume the object bounds to be four independent variables, which could be regressed by the l2 loss separately. Such an oversimplified assumption is contrary to the well-received observation, that those variables are correlated, resulting to less accurate localization. To address the issue, we firstly introduce a novel Intersection over Union (IoU) loss function for bounding box prediction, which regresses the four bounds of a predicted box as a whole unit. By taking the advantages of IoU loss and deep fully convolutional networks, the UnitBox is introduced, which performs accurate and efficient localization, shows robust to objects of varied shapes and scales, and converges fast. We apply UnitBox on face detection task and achieve the best performance among all published methods on the FDDB benchmark.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

FCOS: Fully Convolutional One-Stage Object Detection (2019, methodology): “Another family of anchor-free detectors such as [24] are based on DenseBox [9].”
Generalized Intersection Over Union: A Metric and a Loss for Bounding Box Regression (2019, methodology): “Similarly, for the object detection task, recent works [8, 24] have attempted to directly or indirectly incorporate IoU to better perform bounding box regression.”
Acquisition of Localization Confidence for Accurate Object Detection (2018, methodology): “In the original Fast R-CNN [5] framework, crit is chosen as an smooth-L1 distance of coordinates in log-scale, while in [32], crit is chosen as the − ln(IoU) between two bounding boxes.”
YOLOv4: Optimal Speed and Accuracy of Object Detection (2020, background): “In order to make this issue processed better, some researchers recently proposed IoU loss [90], which puts the coverage of predicted BBox area and ground truth BBox area into consideration.”
Feature Selective Anchor-Free Module for Single-Shot Object Detection (2019, methodology): “UnitBox [36] proposed an Intersection over Union (IoU) loss function for better box regression.”
Conditional DETR for Fast Training Convergence (2021, methodology): “Typical methods include YOLOv1 [30], CornerNet [21], ExtremeNet [50], CenterNet [49, 6], FCOS [39] and others [23, 28, 52, 19, 51, 22, 15, 46, 47].”
A Survey of Deep Learning-Based Object Detection (2019, background): “A typical work [81] adopts IoU loss to measure the degree of accuracy the network predicting.”
IoU Loss for 2D/3D Object Detection (2019, methodology): “In [1] and [2], the effectiveness of IoU as loss function has been well proved for 2D axis aligned BBox regression task.”
Focal and Efficient IOU Loss for Accurate Bounding Box Regression (2021, methodology): “The IOU loss [25] for measuring similarity between two arbitrary shapes (volumes) A,B ⊆ S ∈ R is attained by: LIOU = 1− |A ∩B| |A ∪B| , (3)”
AOD-Net: All-in-One Dehazing Network (2017, background): “We study the problem of object detection and recognition [16, 30] in the presence of haze, as an example for how high-level vision tasks can interact with dehazing.”
Training-Time-Friendly Network for Real-Time Object Detection (2019, background): “2015) is the first anchor-free detector, and then UnitBox (Yu et al. 2016) upgrades DenseBox for better performance.”
SiamCAR: Siamese Fully Convolutional Classification and Regression for Visual Tracking (2019, background): “(4) where LIOU is the IOU loss as in [38] and I(·) is an indicator function defined by:”
Soft Anchor-Point Object Detection (2019, methodology): “IoU loss [30] (lIoU ) is used for the training of localization subnets.”
Dense Regression Network for Video Grounding (2020, methodology): “where we use the IoU regression loss [41] as L1 following Tian et al.”
Cascade RPN: Delving into High-Quality Region Proposal Network with Adaptive Convolution (2019, methodology): “Rather than relying on anchors, there have been studies that define positive samples by the distance between the prediction points and the center region of objects, referred to as anchor-free [35, 40, 42].”
Object Detection in 20 Years: A Survey (2019, positive): “The most intuitive improvement is to directly use IoU as the localization loss [105], [106], [107], [190].”
SiamFC++: Towards Robust and Accurate Visual Tracking with Target Estimation Guidelines (2019, methodology): “2017) for classification result, Lquality denote the binary cross entropy (BCE) loss for quality assessment and Lreg denote the IoU loss (Yu et al. 2016) for bounding box result.”
FOTS: Fast Oriented Text Spotting with a Unified Network (2018, methodology): “As for the regression loss, we adopt the IoU loss in [52] and the rotation angle loss in [53], since they are robust to variation in object shape, scale and orientation:”