Vision-Based Traffic Sign Detection and Analysis for Intelligent Driver Assistance Systems: Perspectives and Survey

By Andreas Møgelmose, Mohan Manubhai Trivedi, Thomas Baltzer Moeslund. 2012

713 citations recorded by the source corpus.

Subfields: Automotive Engineering, Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

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

In Automotive Engineering

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Robust Physical-World Attacks on Deep Learning Models (2017, methodology): “Therefore, we use the LISA dataset [18], a U.”
Threat of Adversarial Attacks on Deep Learning in Computer Vision: A Survey (2018, methodology): “The classifier is trained using LISA dataset for road signs [174].”
Deep Learning for Image and Point Cloud Fusion in Autonomous Driving: A Review (2020, background): “Existing reviews [93] [100] have comprehensively covered traditional traffic sign”
Computer vision in automated parking systems: Design, implementation and challenges (2017, background): “earlier, it is also important to be able to detect other road markings, such as arrows and disabled parking markings [28], and road signs [29], which allows the autonomously parking vehicle to follow the rules defined in the parking area.”
Vision-Based Driver Assistance Systems: Survey, Taxonomy and Advances (2015, background): “Cameras have become essential sensors in standard alert type functions such as traffic sign recognition (TSR) [5][6], cross traffic alert (CTA) [7], lane departure warning (LDW) [8] and object detection (OD) [7].”
Towards Universal Object Detection by Domain Attention (2019, methodology): “Both KITTI [9] and LISA [33] depict traffic scenes, collected with cameras mounted on moving vehicles.”
Robust Physical-World Attacks on Deep Learning Visual Classification (2018, background): “traffic sign dataset containing 47 different road signs [21].”
Robust Physical-World Attacks on Machine Learning Models (2017, methodology): “Therefore, we used the LISA dataset [14], a U.”
Detection of traffic signs in real-world images: The German traffic sign detection benchmark (2013, background): “[17] provides detailed analysis on the most recent developments .”
Review of Artificial Intelligence Adversarial Attack and Defense Technologies (2019, methodology): “Road Sign Recognition Eykholt et al. [35] proposed a white-box adversarial sample generation method to attack their own trained road sign recognition models, including LISA-CNN models used LISA [81], a U.S. traffic sign dataset containing 47 different road signs, and GTSRB-CNN models, which trained ”
CURE-TSR: Challenging Unreal and Real Environments for Traffic Sign Recognition (2017, background): “It is a common practice to test the robustness of these systems with traffic datasets [1, 2, 3, 4, 5, 6, 7, 8, 9, 10].”
Traffic sign recognition — How far are we from the solution? (2013, background): “No other dataset surveyed in [5] has a comparable size or number of classes.”
Traffic Sign Recognition With Hinge Loss Trained Convolutional Neural Networks (2014, background): “The research around this issue has long been popular [1], but the task is challenging.”
Standard detectors aren't (currently) fooled by physical adversarial stop signs (2017, methodology): “LISA-CNN consists of three convolutional layers followed by a fully connected layer ( [3], p5, c1), trained to classify signs into 17 classes ( [3], p4, c2), using the LISA dataset of US road signs [13].”
Cascaded Segmentation-Detection Networks for Text-Based Traffic Sign Detection (2018, background): “However, in the past, most of the traffic sign detection algorithms mainly deal with symbol-based traffic signs [3]–[6], such as pedestrian crossing, no stopping, speed limit sign, no parking and so on, which have specific shapes and colors”
Latent Backdoor Attacks on Deep Neural Networks (2019, methodology): “The Teacher dataset GTSRB [46] contains 39,200 colored training images and 12,600 testing images, while the Student dataset LISA [35] has 3700 training images of 17 US traffic signs2.”