Hybrid channel based pedestrian detection

Tesema, Fiseha B.; Wu, Hong; Chen, Mingjian; Lin, Junpeng; Zhu, William; Huang, Kaizhu

doi:10.1016/j.neucom.2019.12.110

Cited by 21 publications

(15 citation statements)

References 56 publications

(141 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By employing an attention network to the baseline faster RCNN, better detection was obtained [26]. Tesema et al [27] extended the RPN (Region Proposal Network) + BF (Boosted Forest) framework [28] with handcrafted features and CNN features. ROI pooling is applied both on handcrafted features and CNN features for pedestrian candidate proposal.…”

Section: Deep Learning-based Pedestrian Detectionmentioning

confidence: 99%

Pedestrian Detection Based on Two-Stream UDN

Wang

et al. 2020

Applied Sciences

View full text Add to dashboard Cite

Pedestrian detection is the core of the driver assistance system, which collects the road conditions through the radars or cameras on the vehicle, judges whether there is a pedestrian in front of the vehicle, supports decisions such as raising the alarm, automatically slowing down, or emergency stopping to keep pedestrians safe, and improves the security when the vehicle is moving. Suffering from weather, lighting, clothing, large pose variations, and occlusion, the current pedestrian detection still has a certain distance from the practical applications. In recent years, deep networks have shown excellent performance for image detection, recognition, and classification. Some researchers employed deep network for pedestrian detection and achieve great progress, but deep networks need huge computational resources, which make it difficult to put into practical applications. In real scenarios of autonomous vehicles, the computation ability is limited. Thus, the shallow networks such as UDN (Unified Deep Networks) is a better choice, since it performs well while consuming less computation resources. Based on UDN, this paper proposes a new deep network model named two-stream UDN, which augments another branch for solving traditional UDN’s indistinction of the difference between trees/telegraph poles and pedestrians. The new branch accepts the upper third part of the pedestrian image as input, and the partial image has less deformation, stable features, and more distinguished characters from other objects. For the proposed two-stream UDN, multi-input features including the HOG (Histogram of Oriented Gradients) feature, Sobel feature, color feature, and foreground regions extracted by GrabCut segmentation algorithms are fed. Compared with the original input of UDN, the multi-input features are more conducive for pedestrian detection, since the fused HOG features and significant objects are more significant for pedestrian detection. Two-stream UDN is trained through two steps. First, the two sub-networks are trained until converge; then, we fuse results of the two subnets as the final result and feed it back to the two subnets to fine tune network parameters synchronously. To improve the performance, Swish is adopted as the activation function to obtain a faster training speed, and positive samples are mirrored and rotated with small angles to make the positive and negative samples more balanced.

show abstract

Section: Deep Learning-based Pedestrian Detectionmentioning

confidence: 99%

Pedestrian Detection Based on Two-Stream UDN

Wang

et al. 2020

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…However, people can suffer from occlusion as well as variations in illumination, scale, and background, which make human detection in indoor scenes a challenging task. Methods based only on RGB features [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 ] can no longer meet the needs of human detection in many scenarios. With the popularization of inexpensive depth acquisition equipment, detecting human with the help of depth information has become an effective and feasible scheme.…”

Section: Introductionmentioning

confidence: 99%

“…To address the challenges of occlusion and scale changes in RGB images, several pedestrian detection algorithms [ 7 , 8 , 9 , 10 , 11 , 12 , 13 ] have been developed based on novel processing approaches. Andre et al [ 7 ] proposed a cascaded aggregate channel features (ACF) detector to accurately detect humans.…”

Section: Introductionmentioning

confidence: 99%

“…Cheng et al [ 8 ] divided the human body into multiple body parts and introduced a fast fused part-based model (FFPM) for pedestrian detection in a crowded environment. Tesema et al [ 9 ] extended the Faster R-CNN framework to combine handcrafted features and CNN features for pedestrian detection. Wang et al [ 10 ] introduced a RepLoss loss function to detect humans.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Asymmetric Adaptive Fusion in a Two-Stream Network for RGB-D Human Detection

Zhang

Guo

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

In recent years, human detection in indoor scenes has been widely applied in smart buildings and smart security, but many related challenges can still be difficult to address, such as frequent occlusion, low illumination and multiple poses. This paper proposes an asymmetric adaptive fusion two-stream network (AAFTS-net) for RGB-D human detection. This network can fully extract person-specific depth features and RGB features while reducing the typical complexity of a two-stream network. A depth feature pyramid is constructed by combining contextual information, with the motivation of combining multiscale depth features to improve the adaptability for targets of different sizes. An adaptive channel weighting (ACW) module weights the RGB-D feature channels to achieve efficient feature selection and information complementation. This paper also introduces a novel RGB-D dataset for human detection called RGBD-human, on which we verify the performance of the proposed algorithm. The experimental results show that AAFTS-net outperforms existing state-of-the-art methods and can maintain stable performance under conditions of frequent occlusion, low illumination and multiple poses.

show abstract

“…Recently, Deep Learning models (CNNs) have achieved great success in object detection [9]- [14]. Naturally, they have also been applied to pedestrian detection by using the learned deep feature to improve the detection performance [15]- [22], [37], [41], [42]. While these pedestrian detection methods using Deep Learning models have improved the detecting accuracy, they demand more complex architectures and higher computational costs, and also need special hardware to run in reasonable time.…”

Section: Introductionmentioning

confidence: 99%

Pedestrian Detection Using Quaternion Gradient Based Weber Local Descriptor

Lian

2021

IEEE Access

View full text Add to dashboard Cite

In the past decades, pedestrian detection has attracted more attention in many practical applications. In this paper, a novel pedestrian detection method using Quaternion Gradient based Weber Local Descriptor (QGWLD) is proposed. Unlike many other local pedestrian detectors that only extracted features from the gray-scale image, which ignored the color information, a new powerful and robust local pedestrian detector is developed, which integrates the advantages of both the color and texture information and can be used to address the pedestrian detection problem well. As we know, the quaternion can entirely characterize a color object well and the WLD feature consisting of the gradient orientation and differential excitation has acquired a good performance for pedestrian detection. Therefore, combining the WLD feature with the quaternion representation, the QGWLD pedestrian detector is presented. Firstly, the quaternion gradient representation is performed on the color image in the sliding window, and then the Weber Local Descriptor (WLD) is extracted over the quaternion gradient feature map. After that, the QGWLD histogram is constructed to characterize the sliding window. Experimental results on INRIA and PennFudanPed pedestrian databases validate the effectiveness of the proposed QGWLD pedestrian detector. Comparing with the similar pedestrian detectors, the proposed QGWLD pedestrian detector performs better.

show abstract

Hybrid channel based pedestrian detection

Cited by 21 publications

References 56 publications

Pedestrian Detection Based on Two-Stream UDN

Pedestrian Detection Based on Two-Stream UDN

Asymmetric Adaptive Fusion in a Two-Stream Network for RGB-D Human Detection

Pedestrian Detection Using Quaternion Gradient Based Weber Local Descriptor

Contact Info

Product

Resources

About