Abstract
Pedestrian detection aims to localize and recognize every pedestrian instance in an image with a bounding box. The current state-of-the-art method is Faster RCNN, which is such a network that uses a region proposal network (RPN) to generate high quality region proposals, while Fast RCNN is used to classifiers extract features into corresponding categories. The contribution of this paper is integrated low-level features and high-level features into a Faster RCNN-based pedestrian detection framework, which efficiently increase the capacity of the feature. Through our experiments, we comprehensively evaluate our framework, on the Caltech pedestrian detection benchmark and our methods achieve state-of-the-art accuracy and present a competitive result on Caltech dataset.
Highlights
Pedestrian detection has been exhaustively explored in the last decade because of its growing importance in realworld applications, such as automatic driving, human behavior analysis [1] or intelligent video surveillance [2]
The most classic method based on convolutional neural network (CNN) is Faster R-CNN, which consists of two components: a region proposal network (RPN) to generate nearly cost-free region proposals, and uses a Fast RCNNclassifier to detect pedestrians, which has shown leading accuracy on several multi-category benchmarks
After that researcher continue to develop new convolution neural network model, such as You Only Look Once (YOLO), Single Shot Multi Box Detector (SSD) and so on, First, typical scenarios of pedestrian detection, such as automatic driving and intelligent surveillance, generally the size of pedestrian instances in small object image are less than 30x40 pixels. this make it difficult for the classifier to identify whether it is a pedestrian or a nonpedestrian
Summary
Pedestrian detection has been exhaustively explored in the last decade because of its growing importance in realworld applications, such as automatic driving, human behavior analysis [1] or intelligent video surveillance [2]. The second approach is through the deep learning technology to achieve the goal of pedestrian detection, and have outperformed state-of-art performance on several pedestrian datasets This method uses the convolutional neural network (CNN) to automatically extract the global and semantic features of the image, generate high-quality candidate boxes and classify each candidate. After that researcher continue to develop new convolution neural network model, such as You Only Look Once (YOLO), Single Shot Multi Box Detector (SSD) and so on, First, typical scenarios of pedestrian detection, such as automatic driving and intelligent surveillance, generally the size of pedestrian instances in small object image are less than 30x40 pixels. The evaluation results show that our network achieves the state-of-the-art performance on Caltech datasets
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
