A Fast Stereo-based System for Detecting and Tracking Pedestrians from a Moving Vehicle

Bajracharya, Max; Moghaddam, B.; Howard, Andrew; Brennan, Shane; Matthies, Larry

doi:10.1177/0278364909341884

Cited by 74 publications

(75 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another recent approach used a fast coarse-to-fine detection scheme but sacrificed detection of small pedestrians [35] (which are of crucial importance in many applications [8,10]). A number of fast systems have been proposed [3,12]; however, a detailed overview is outside of the scope of this work. Finally, a number of groups have recently ported HOG to a parallel implementation using GPUs [4,32,34]; such efforts are complementary to our own.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

The Fastest Pedestrian Detector in the West

Dollár¹,

Belongie²,

Perona³

2010

Procedings of the British Machine Vision Conference 2010

494

361

View full text Add to dashboard Cite

We demonstrate a multiscale pedestrian detector operating in near real time (∼6 fps on 640x480 images) with state-of-the-art detection performance. The computational bottleneck of many modern detectors is the construction of an image pyramid, typically sampled at 8-16 scales per octave, and associated feature computations at each scale. We propose a technique to avoid constructing such a finely sampled image pyramid without sacrificing performance: our key insight is that for a broad family of features, including gradient histograms, the feature responses computed at a single scale can be used to approximate feature responses at nearby scales. The approximation is accurate within an entire scale octave. This allows us to decouple the sampling of the image pyramid from the sampling of detection scales. Overall, our approximation yields a speedup of 10-100 times over competing methods with only a minor loss in detection accuracy of about 1-2% on the Caltech Pedestrian dataset across a wide range of evaluation settings. The results are confirmed on three additional datasets (INRIA, ETH, and TUD-Brussels) where our method always scores within a few percent of the state-of-the-art while being 1-2 orders of magnitude faster. The approach is general and should be widely applicable.

show abstract

Section: Related Workmentioning

confidence: 99%

“…The total cost is 4 3 n 2 , which is only 33% more than the cost of computing single scale features. Typical detectors are evaluated on m = 8 to 16 scales per octave [8], thus according to (3) we expect an order of magnitude savings by using the proposed hybrid approach (more if upsampled images are used).…”

Section: Complexity Analysismentioning

confidence: 99%

The Fastest Pedestrian Detector in the West

Dollár¹,

Belongie²,

Perona³

2010

Procedings of the British Machine Vision Conference 2010

494

361

View full text Add to dashboard Cite

show abstract

“…and motion. While some approaches utilize combinations on the module level [2], [12], [13], [19], [37], [47], others integrate multiple information sources directly into the pattern classification step [5], [9], [40], [43], [44], [49], [51]- [53], [56], [58].…”

Section: Previous Workmentioning

confidence: 99%

A Multilevel Mixture-of-Experts Framework for Pedestrian Classification

Enzweiler

Gavrila

2011

IEEE Trans. on Image Process.

126

View full text Add to dashboard Cite

Abstract-Notwithstanding many years of progress, pedestrian recognition is still a difficult but important problem. We present a novel multilevel Mixture-of-Experts approach to combine information from multiple features and cues with the objective of improved pedestrian classification. On pose-level, shape cues based on Chamfer shape matching provide sample-dependent priors for a certain pedestrian view. On modality-level, we represent each data sample in terms of image intensity, (dense) depth, and (dense) flow. On feature-level, we consider histograms of oriented gradients (HOG) and local binary patterns (LBP). Multilayer perceptrons (MLP) and linear support vector machines (linSVM) are used as expert classifiers.Experiments are performed on a unique real-world multimodality dataset captured from a moving vehicle in urban traffic. This dataset has been made public for research purposes. Our results show a significant performance boost of up to a factor of 42 in reduction of false positives at constant detection rates of our approach compared to a baseline intensity-only HOG/linSVM approach.

show abstract

“…In the field of intelligent vehicles, most stereo-based multiple object tracking methods have been concerned with object detection and localization problems in 3D global coordinates [8][9][10][11]. There has been a lack of efforts to increase the accuracy and precision of the ROI in the image plane.…”

Section: Introductionmentioning

confidence: 99%

Stereo-Based Tracking-by-Multiple Hypotheses Framework for Multiple Vehicle Detection and Tracking

Lim

Kim

Lee

2013

International Journal of Advanced Robotic Systems

View full text Add to dashboard Cite

In this paper, we present a tracking-by-multiple hypotheses framework to detect and track multiple vehicles accurately and precisely. The tracking-bymultiple hypotheses framework consists of obstacle detection, vehicle recognition, visual tracking, global position tracking, data association and particle filtering. The multiple hypotheses are from obstacle detection, vehicle recognition and visual tracking. The obstacle detection detects all the obstacles on the road. The vehicle recognition classifies the detected obstacles as vehicles or non-vehicles. 3D feature-based visual tracking estimates the current target state using the previous target state. The multiple hypotheses should be linked to corresponding tracks to update the target state. The hierarchical data association method assigns multiple tracks to the correct hypotheses with multiple similarity functions. In the particle filter framework, the target state is updated using the Gaussian motion model and the observation model with associated multiple hypotheses. The experimental results demonstrate that the proposed method enhances the accuracy and precision of the region of interest.

show abstract

A Fast Stereo-based System for Detecting and Tracking Pedestrians from a Moving Vehicle

Cited by 74 publications

References 16 publications

The Fastest Pedestrian Detector in the West

The Fastest Pedestrian Detector in the West

A Multilevel Mixture-of-Experts Framework for Pedestrian Classification

Stereo-Based Tracking-by-Multiple Hypotheses Framework for Multiple Vehicle Detection and Tracking

Contact Info

Product

Resources

About