Pedestrian detection at 100 frames per second

By Rodrigo Benenson, Markus Mathias, Radu Timofte, Luc Van Gool. 2012

621 citations recorded by the source corpus.

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

DeepSetNet: Predicting Sets with Deep Neural Networks (2016, background): “Traditional approaches follow the slidingwindow paradigm [44, 7, 46, 11, 2], where each possible (or rather plausible) image region is scored independently to contain a person or not.”
The Cityscapes Dataset for Semantic Urban Scene Understanding (2016, background): “Specialized object-centric methods fall somewhere in between by focusing on detecting a certain subset of (mostly dynamic) scene constituents, e.g. [6,12,13,15].”
A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection (2016, background): “2 (c) is to rescale the input a few times and learn a small number of model templates [24].”
Deep Feature Pyramid Reconfiguration for Object Detection (2018, background): “Several attempts have been performed on image pyramids for the sake of eﬃcient computation [4,7,8].”
Finding Tiny Faces (2016, background): “We provide an in-depth exploration of scale-variant templates, which have been previously proposed for pedestrian detection[17], sometimes in the context of improved speed [3].”
Local Decorrelation For Improved Pedestrian Detection (2014, background): “Recent boosted detectors [1] [7] achieve state-of-the-art accuracy on modern benchmarks [10] [22] while retaining computational efficiency.”
RIGOR: Reusing Inference in Graph Cuts for Generating Object Regions (2014, background): “When enumerating bounding boxes and computing their features, tricks such as integral image and distance transform [10, 23] can help to speed-up the computation, even achieving real-time performance in certain tasks [3, 29].”
SAN: Learning Relationship between Convolutional Features for Multi-Scale Object Detection (2018, background): “Multi-scale objects are detected by a single detector [5,7,8,33] that uses an image pyramid with scale normalization, or by a multi-scale detector [2] that uses a separate detector for each of several scales.”
Detection of traffic signs in real-world images: The German traffic sign detection benchmark (2013, methodology): “Recently, variants of the integral channel features detector have shown excellent results in terms of speed (50 fps) and quality for the task of pedestrian detection, improving over most existing methods [35], [36].”
Regionlets for Generic Object Detection (2013, methodology): “Object detection is arguably an indispensable component for most vision tasks, and it has achieved prominent success for specific targets such as faces [ 1], [18] and pedestrians [3], [19], [20], [21], [22], [23], [24].”
Joint Human Detection and Head Pose Estimation via Multistream Networks for RGB-D Videos (2017, methodology): “[4] support efficient and accurate pedestrian detection with the proposed methods reaching speeds of 100 to 135 fps.”
Deep Learning Strong Parts for Pedestrian Detection (2015, background): “Pedestrian detection has been studied extensively in recent years [4, 7, 6, 1, 2, 29, 42] and has many applications such as video surveillance and robotics.”
Scale-Aware Fast R-CNN for Pedestrian Detection (2015, methodology): “Hand-crafted Model: The models based on hand-crafted features have been widely used for object detection [2], [5], [8]–[10], [13], [18], [19], [30], [39], [51], [54], [61], [62].”
Pedestrian detection with a Large-Field-Of-View deep network (2015, methodology): “Fast methods, such as WordChannels [40] or VeryFast [7], which run at more than 16 frames per second of GPU or CPU, are at the same time not among the best performing.”
Face Detection without Bells and Whistles (2014, background): “By using aggressive (soft and crosstalk) cascades and reducing features computation across scales [2,6], it has been shown that integral channels detectors can reach fast detection rates of the order of ∼ 50 Hz for 640 × 480 pixels images (either on GPU or multi-core CPU).”
Ten Years of Pedestrian Detection, What Have We Learned? (2014, background): “This technique does not impact computational cost at detection time [53], although training time increases.”
Traffic sign detection and recognition using fully convolutional network guided proposals (2016, background): “Previous works usually adopt coarse-to-ﬁ ne strategies which help to signi ﬁ cantly speed up object detection process [50,51].”
Object Detection in 20 Years: A Survey (2019, background): “, to detect objects of different scales by simply sliding multiple detectors on one feature map rather than re-scaling the image or features [173].”