Gaussian YOLOv3: An Accurate and Fast Object Detector Using Localization Uncertainty for Autonomous Driving

By Jiwoong Choi, Dayoung Chun, Hyun Kim, Hyuk-Jae Lee. 2019

457 citations recorded by the source corpus.

The use of object detection algorithms is becoming increasingly important in autonomous vehicles, and object detection at high accuracy and a fast inference speed is essential for safe autonomous driving. A false positive (FP) from a false localization during autonomous driving can lead to fatal accidents and hinder safe and efficient driving. Therefore, a detection algorithm that can cope with mislocalizations is required in autonomous driving applications. This paper proposes a method for improving the detection accuracy while supporting a real-time operation by modeling the bounding box (bbox) of YOLOv3, which is the most representative of one-stage detectors, with a Gaussian parameter and redesigning the loss function. In addition, this paper proposes a method for predicting the localization uncertainty that indicates the reliability of bbox. By using the predicted localization uncertainty during the detection process, the proposed schemes can significantly reduce the FP and increase the true positive (TP), thereby improving the accuracy. Compared to a conventional YOLOv3, the proposed algorithm, Gaussian YOLOv3, improves the mean average precision (mAP) by 3.09 and 3.5 on the KITTI and Berkeley deep drive (BDD) datasets, respectively. Nevertheless, the proposed algorithm is capable of real-time detection at faster than 42 frames per second (fps) and shows a higher accuracy than previous approaches with a similar fps. Therefore, the proposed algorithm is the most suitable for autonomous driving applications.

Subfields: Aerospace Engineering, 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 Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection (2020, background): “Recently, Gaussian assumption [10, 4] is adopted to learn the uncertainty by introducing a predicted variance.”
YOLOv4: Optimal Speed and Accuracy of Object Detection (2020, methodology): “We also conduct experiments on various BoS, including Mish, SPP, SAM, RFB, BiFPN, and Gaussian YOLO [8].”
A Review and Comparative Study on Probabilistic Object Detection in Autonomous Driving (2020, methodology): “Nevertheless, the majority of recent work on probabilistic object detection [9], [12], [14]–[17], [88], [90], [91], [94] still use mAP as the only metric to provide a quantitative assessment of their proposed methods, emphasizing a secondary eﬀect of accuracy improvement when integrating…”
Dynamic Anchor Learning for Arbitrary-Oriented Object Detection (2020, methodology): “Regression uncertainty has been widely discussed in many previous work (Feng, Rosenbaum, and Di-etmayer 2018; Choi et al. 2019; Kendall and Gal 2017; Choi et al. 2018), which represents the instability and irrelevance in the regression process.”
Edge YOLO: Real-Time Intelligent Object Detection System Based on Edge-Cloud Cooperation in Autonomous Vehicles (2022, methodology): “Among those algorithms, the most representative one is the You Only Look Once (YOLO) [13] series of object detection networks.”
A survey of uncertainty in deep neural networks (2021, background): “medical image analysis [1], [2], [3], [4], [5], [6] or autonomous vehicle control [7], [8], [9], [10], their deployment in mission-”
Imbalance Problems in Object Detection: A Review (2019, methodology): “The idea is similar to the networks with an additional localization confidence associated prediction branch [89], [125], [126], [127], [128], besides classification and regression, in order to use the predicted confidence during inference.”
Localization Distillation for Object Detection (2022, background): “[53], [56] to Gaussian distribution [9], [23], and further to probability distribution [38], [51].”
MonoPair: Monocular 3D Object Detection Using Pairwise Spatial Relationships (2020, background): “Gaussian-YOLO [9] significantly improves the detection results by predicting the localization uncertainty.”
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,”
PP-YOLO: An Effective and Efficient Implementation of Object Detector (2020, methodology): “Anchor-based methods are still the mainstream of object detection [33, 21, 31, 32, 1, 22, 28, 9, 45, 2, 5, 37, 20, 4, 15], which evolved from early proposal based detectors, such as Fast R-CNN [11].”
A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges (2020, background): “[153] proposed a Gaussian-based model to predict the localization uncertainty in YOLOv3 [154].”
Data Uncertainty Learning in Face Recognition (2020, background): “ameters in deep neural networks, and data uncertainty measuring the noise inherent in given training data. Recently, many computer vision tasks, i.e., semantic segmentation [19, 21], object detection [6, 25] and person Re-ID [50], have introduced deep uncertainty learning to CNNs for the improvement”
Probabilistic Anchor Assignment with IoU Prediction for Object Detection (2020, background): “In contrast to predicting IoUs, some works [4, 12] add an additional head to predict the variance of localization to regularize training [12] or penalize the classification score in testing [4].”
Uncertainty-Aware Learning for Zero-Shot Semantic Segmentation (2020, background): “[9] extend object detection frameworks to be probabilistic and learn to estimate the uncertainty of detection results.”
Prime-Aware Adaptive Distillation (2020, background): “, face recognition [30], semantic segmentation [15,16], object detection [7,17] and person re-identification(re-ID) [38].”