Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-data training detection model generation method, system and device and storage medium

A detection model and multi-data technology, applied in neural learning methods, biological neural network models, instruments, etc., can solve problems such as lack of data target detection model, insufficient data set association transition, etc., to achieve the effect of rich expression ability

Pending Publication Date: 2022-07-05
人民中科(北京)智能技术有限公司
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention provides a detection model generation method, system, equipment and storage medium for multi-data training, which are used to solve the problem that the prior art cannot associate sufficient data sets to deficient data sets, and then train to obtain a target detection model with deficient data as the task The problem

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
  • Multi-data training detection model generation method, system and device and storage medium
  • Multi-data training detection model generation method, system and device and storage medium
  • Multi-data training detection model generation method, system and device and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] Embodiment 1. For the detection model generation method of multi-data training in this embodiment, see figure 1 shown, including the following main steps:

[0045] 100. Prepare pre-training data.

[0046] In order to improve the expressive ability of pre-training, in this embodiment, both the classification set data and the detection set data are used for training, and the corresponding relationship between different categories and the balanced sampling of the training data also need to be considered during training. When preparing the pre-training data, the first classification data set, the first detection data set and the second detection data set are used, and differential sampling is performed according to the samples in the first classification data set and the first detection data set, and the categories with a large number of samples are selected. Use a low sampling probability, and use a high sampling probability for a category with a small number of samples, ...

Embodiment 2

[0083] Embodiment 2. For the detection model generation system for multi-data training in this embodiment, see Figure 5 As shown, it includes: preparing pre-training data unit 200 , first input unit 210 , output unit 220 , pre-training detection network model unit 230 , second input unit 240 , generating unit 250 , and mapping unit 260 .

[0084] The pre-training data unit 200 is used for preparing the pre-training data, using the first classification data set, the first detection data set and the second detection data set, and according to the first classification data set, the first detection data set Differential sampling is performed on the samples of the sample, and the category with a large number of samples uses a low sampling probability, and the category with a small number of samples uses a high sampling probability. Not the same, some have only a few hundred sample images, and some are counted as 100,000 sample images. Different sampling rates are used for sample i...

Embodiment 3

[0092] Embodiment 3, the computer equipment of this embodiment, see Image 6 As shown, the computer device 300 shown is only an example, and should not impose any limitation on the function and scope of use of the embodiments of the present invention.

[0093] like Image 6 As shown, computer device 300 takes the form of a general-purpose computing device. Components of computer device 300 may include, but are not limited to, one or more processors or processing units 301, system memory 302, and a bus 303 connecting various system components including system memory 302 and processing unit 301.

[0094] Bus 303 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of a variety of bus structures. By way of example, these architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bu...

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 multi-data training detection model generation method, system and device and a storage medium. The method comprises the following steps: inputting a pre-training data set into a basic network in a matrix data mode; the basic network outputs coordinates and target labels of a plurality of target bounding boxes; according to the output coordinates of the plurality of target bounding boxes and the target labels, loss of the basic network is obtained, model pre-training of the basic network is completed, and a pre-training detection network model is formed; inputting the task data set into a pre-training detection network model; and adjusting the pre-training detection network model according to the task data set, and generating a multi-data training detection model. The system comprises a first input unit, an output unit, a pre-training detection network model unit, a second input unit and a generation unit. A computer device includes a memory, a processor, and a computer program. The storage medium comprises computer executable instructions and is used for executing the method.

Description

technical field [0001] The present invention relates to the technical field of computer machine vision, in particular to a method, system, device and storage medium for generating a target detection model based on multi-data joint training. Background technique [0002] The rapid development of deep learning techniques and large-scale categories of labeled data have driven the development and advancement of computer vision tasks, including image recognition, object detection, and image segmentation. Among them, object detection technology has received extensive attention as a basic task with a wide range of applications and needs in computer vision tasks. [0003] In real tasks, many datasets have the problem of lack of data, such as a classification problem of endangered birds. Since they are endangered birds, they are naturally some rare species, so it is impossible for these birds to exist in a large number of various types. , but considering the high similarity between ...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/764G06V10/774G06V10/82G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/2415G06F18/214
Inventor 陈文晶王坚李兵余昊楠胡卫明
Owner 人民中科(北京)智能技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com