Target distribution detection method and equipment
A distribution detection and distribution prediction technology, applied in neural learning methods, instruments, biological neural network models, etc., can solve the problem of low accuracy, and achieve the effect of improving accuracy, expressive ability and robust performance
Inactive Publication Date: 2018-11-23
北京飞搜科技有限公司
View PDF4 Cites 9 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the accuracy rate of the target distribution detection scheme in the prior art is not high
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0030] Research on the existing target distribution detection scheme found that the existing target distribution detection methods are as follows:
[0031] a. Crowd density analysis method based on statistical features Perform crowd density analysis on the input video, obtain the crowd density value of a single monitoring point in real time, and then realize the mutual conversion between crowd density and number of people through multi-segment linear fitting method. The disadvantage is that the feature extraction method here is relatively traditional and the accuracy is not high.
[0032] b. Determine the combined form of the binary classifier for the region in the video image sample, analyze and select the confidence training samples and train each binary classifier separately, and obtain the density level that maximizes the posterior probability with the help of the channel transmission model. The disadvantage is that the system is complex and the method is complex. In some ...
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
Login to View More Abstract
The invention provides a target distribution detection method and equipment. According to the method, a training image is processed based on a convolutional neural network to obtain a target distribution prediction value in the training image, wherein the target distribution prediction value comprises at least one of a target distribution density prediction value and a target quantity category prediction value; a first difference value of a target distribution true value and the target distribution prediction value in the training image is acquired; and parameters of the convolutional neural network are determined according to the first difference value, and a convolutional neural network model is obtained. In this way, the convolutional neural network model can be utilized to perform actual target distribution detection. By utilizing the training image to train the convolutional neural network, the feature representation ability and robustness of the convolutional neural network are better, and the accuracy of target distribution prediction through the convolutional neural network model can be improved.
Description
technical field [0001] The present application relates to the field of Internet information processing technology and computer technology, and in particular to a target distribution detection method and equipment. Background technique [0002] In some scenarios, it is necessary to monitor the distribution of objects, crowds and other targets. For example, crowd numbers and densities have recently received increasing attention as the world's population grows exponentially and urbanization brings increased crowd gathering events such as sporting events, political rallies, and public speaking. In this context, it is very necessary to analyze crowd images for management and security needs. [0003] The prior art monitors the distribution of targets, usually by processing the frame images in the surveillance video, and predicting parameters such as the quantity or distribution density of the targets according to the features of the targets detected in the frame images. However,...
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
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06N3/04G06N3/08
CPCG06N3/08G06V20/53G06N3/045
Inventor 杨旭董远白洪亮熊风烨
Owner 北京飞搜科技有限公司