Training sample acquisition method and device, equipment and computer storage medium

A technology for training samples and obtaining units, which is applied in the field of Internet technology applications, can solve problems such as high labor costs, poor performance of image recognition algorithms, and long algorithm development cycles, and achieve the effect of reducing labor costs and simplifying the acquisition process
CN107909088AActive Publication Date: 2018-04-13BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
20 Cites 9 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
Publication Date
2018-04-13

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention provides a training sample acquisition method and device, equipment and a computer storage medium. The training sample acquisition method comprises the steps of acquiring a first pictureset of a labeled object; searching the first picture set by using a network search engine to obtain a second picture set; integrating the first picture set and the second picture set, enabling the object to serve as a labeling result corresponding to a picture set acquired by integration; and enabling the picture set acquired by integration and the labeling result to serve as training samples ofan image recognition model. by using the technical scheme provided by the invention, a lot of pictures are acquired through the network search engine to serve as training samples, thereby not relyingon the manpower to perform shooting and labeling, thus reducing the manpower cost, and simplifying the acquisition process of the training samples.
Need to check novelty before this filing date? Find Prior Art

Description

【Technical field】

[0001] The present invention relates to the application of Internet technology, in particular to a method, device, equipment and computer storage medium for acquiring training samples. 【Background technique】

[0002] In the field of computer vision, the collection and calibration of image data is the earliest step in all image recognition algorithms. Therefore, the amount of image training data directly affects the quality of the entire image recognition algorithm. Under normal circumstances, researchers need to collect a large amount of image data in the early stage of image recognition algorithm development. In the existing technology, the positive and negative samples are generally collected manually, mainly after the user takes enough pictures of the object, and then performs manual calibration. However, since the image data needs to cover all possible angles and postures of the object, and each image data needs to be calibrated manually, the algorith...

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