Pedestrian detection method based on self-learning

A pedestrian detection and self-learning technology, applied in the field of pedestrian detection based on self-learning, can solve problems such as inability to adapt to specific tasks

Active Publication Date: 2017-06-13
HEFEI NORMAL UNIV
View PDF4 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Purpose of the invention: In order to overcome the inability to adapt to specific tasks in the prior art, provide a pedestrian detection me...

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Pedestrian detection method based on self-learning
  • Pedestrian detection method based on self-learning
  • Pedestrian detection method based on self-learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0053] Such as figure 1 As shown, the present invention provides a pedestrian detection method based on self-learning, comprising the following specific steps: first train a cascade classifier based on AdaBoost as an offline classifier, and simultaneously use a group of public pedestrian photos to train a Gaussian mixture model, feature encoding Using HOG features and location information, and then using a low-threshold offline classifier to detect pedestrians in specific scenes, and output the confidence scores of candidate objects, then select high confidence scores as positive samples, low confidence scores as negative samples, and use Gaussian The hybrid model re-represents the candidate detection objects, and finally uses the SVM classifier to train a discriminative pedestrian classifier online to re-predict the candidate objects and output probability estimates.

[0054] The offline classifier described in the above steps: the training data comes from any pedestrian data...

Embodiment 2

[0059] Such as figure 2 As shown, the offline classifier is trained using the cascade classifier in OpenCV, including two parts: data set preparation and running training program, wherein the data set is prepared as training data: use opencv_createsamples to create a set of positive samples, and manually prepare a large number of negative samples. Sample picture; run the training program to train the cascade classifier: set the feature type to LBP. Considering that the training and detection speed of the LBP feature is several times faster than that of the Haar feature, the LBP feature is used to train the cascade classifier using opencv_traincascade.

Embodiment 3

[0061] Such as image 3 As shown, the Gaussian mixture model includes two parts of feature encoding and GMM training, specifically:

[0062] Feature encoding: For each pedestrian picture in the INRIA dataset, a three-layer Gaussian pyramid is first constructed, and then overlapping image blocks are extracted from each layer of the image pyramid. Assume that each pedestrian picture includes N image blocks Extract the HOG feature hog of each image block pi and its location information l pi =[xy] T , and finally the feature encoding of each image block is f pi =[hog pi T , l pi T ] T , all image blocks constitute pedestrian sample features

[0063] GMM training: A Gaussian mixture model trained offline can be expressed as

[0064]

[0065] Among them, K is the number of Gaussian mixture components, I is the identity matrix, is the mixture weight composed of the kth Gaussian, is a Gaussian distribution with mean μ k and the variance is f is the pedestrian...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a pedestrian detection method based on self-learning. The method comprises the following specific steps that: firstly, training an AdaBoost-based cascade classifier as an off-line classifier, meanwhile, using one group of public pedestrian photos to train a Gaussian mixture model, and adopting HOG (Histogram of Oriented Gradient) feature and position information for feature coding; then, adopting the off-line classifier of a low threshold value to carry out pedestrian detection on a specific scene, and outputting the confidence score of a candidate object; then, picking up a high confidence score as a positive sample and a low confidence score as a negative sample, and using a Gaussian mixture model to show the candidate detection object again; and finally, using a SVM (Support Vector Machine) classifier to train a pedestrian classifier with discriminating ability on line, predicting the candidate object again, and estimating an output probability. By use of the method, the problem that a traditional pedestrian detection method can not carry out adaptation on a specific scene is solved, and the method has a certain promoting effect on a pedestrian detection technology under the specific scene. Compared with a traditional pedestrian detection method, the pedestrian detection is characterized in that a recognition rate is obviously improved.

Description

technical field [0001] The invention relates to a pedestrian detection method in the field of intelligent transportation, in particular to a pedestrian detection method based on specific scene self-learning. Background technique [0002] The problem of road traffic safety has seriously affected economic development and social construction. Reducing the occurrence of road traffic accidents and casualties is an important issue related to people's livelihood. Road traffic safety issues are affected by multiple factors such as pedestrians, vehicles, and roads. Since pedestrians are the main participants and vulnerable persons in road traffic, ensuring pedestrian safety is the key to road safety issues and an important aspect in the field of intelligent transportation systems. Task. [0003] The pedestrian detection method is the core supporting technology of the intelligent transportation system, which has a profound impact on ensuring the safety of pedestrians and reducing the...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06K9/00G06K9/62
CPCG06V40/103G06F18/2411
Inventor 施培蓓曹风云胡玉娟杨雪洁王璐钱言玉王筱薇倩张娜谢超吴友情
Owner HEFEI NORMAL UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap