Target detection method and device based on sparse convolutional neural network
A convolutional neural network and target detection technology, applied in the field of target detection methods and detection devices based on sparse convolutional neural networks, can solve the problems of task redundancy, low detection accuracy of small objects, affecting detection speed, etc., and achieve accurate The effect of target detection
Pending Publication Date: 2020-12-15
上海创屹科技有限公司
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The existing object detection method based on convolutional neural network has low detection accuracy for small objects in large scenes, and its neural network structure has more redundancy for specific tasks, which seriously affects the detection speed
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
[0047] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the following will clearly illustrate the spirit of the content disclosed in the application with the accompanying drawings and detailed descriptions. After any person skilled in the art understands the embodiments of the content of the application , when it can be changed and modified by the technology taught in the content of the application, it does not depart from the spirit and scope of the content of the application.
[0048]The exemplary embodiments and descriptions of the present application are used to explain the present application, but not to limit the present application. In addition, elements / members with the same or similar numbers used in the drawings and embodiments are used to represent the same or similar parts.
[0049] The terms "first", "second", ... etc. used herein do not specifically refer to a sequence or order, nor are they ...
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 detection method and detection device based on a sparse convolutional neural network, and the method comprises the steps of building a data set through employing an obtained target scene image, and marking all detection targets contained in each image in the data set; optimizing the structure of the reference convolutional neural network according to the proportionrange of the detection target in the image to obtain an optimized convolutional neural network; performing pre-training of adding a sparse regularization item to a loss function of the optimized convolutional neural network on the constructed data set to obtain a convolutional neural network; performing convolutional neural network sparsification on the convolutional neural network to obtain a sparsified convolutional neural network; carrying out fine tuning training on the sparse convolutional neural network on the data set to obtain a final sparse convolutional neural network; and detectingthe target by using the final sparse convolutional neural network. According to the invention, rapid and accurate target detection can be carried out on small objects in a large scene on cheap hardware.
Description
technical field [0001] The application belongs to the technical field of target detection, and is particularly suitable for detecting small objects in large scenes, and in particular relates to a target detection method and detection device based on a sparse convolutional neural network. Background technique [0002] Existing fast object detection algorithms based on convolutional neural networks are usually implemented by transfer learning training of benchmark models (such as Yolo models) on custom datasets. Among them, the convolutional neural network benchmark model has a complex deep structure. From the perspective of detection targets, most of the targets detected are targets of different scales, and the targets cannot be too small in the image; in terms of detection speed, if it is less than 5ms Most of the ultra-high-speed detection has very high requirements on hardware. The existing object detection method based on convolutional neural network has low detection ac...
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 IPC IPC(8): G06K9/00G06N3/04G06N3/08
CPCG06N3/08G06V20/20G06V2201/07G06N3/045
Inventor 贺琪欲张海波杨跞许楠张文
Owner 上海创屹科技有限公司