A classification and division method of image data set
A technology of image data sets and image data, applied in still image data clustering/classification, still image data retrieval, complex mathematical operations, etc., can solve problems such as defects, inaccurate division, and inability to effectively reduce hash collisions, etc. Achieve the effect of reducing the probability of hash collision and accurate division
Active Publication Date: 2022-07-05
成都成信高科信息技术有限公司 +1
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Although traditional feature learning such as SIFT and HOG can extract some features of images, they have also achieved good results in image classification, but this artificial design feature method has certain defects.
The existing image classification technology is not accurate enough to divide the image data set, and when using the hash algorithm for image data processing, it cannot effectively reduce the probability of hash collision
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
[0023] In order to have a clearer understanding of the technical features, objects and effects of the present invention, the specific embodiments of the present invention will now be described with reference to the accompanying drawings.
[0024] When classifying the image data set, the result of the classification must be strictly that each unit belongs to a certain category, and it is not allowed to belong to this category, to another category, or to be omitted. If images are classified as people, cats, dogs, tables, etc., the images in each class strictly belong to this class and not to other classes, and the sum of the number of images in all classes is equal to the sum of the overall images.
[0025] For images that have been classified and labeled, they can be directly divided. In this embodiment, the present invention mainly proposes a classification and division method of image data sets for images without classification labels, such as figure 1 As shown, the method in...
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 discloses a method for classifying and dividing image data sets, comprising: building a pre-training model, generating a projection matrix group R, scaling picture image data, and importing it into the pre-training model; Perform feature extraction to generate a certain-dimensional feature vector x; for each generated feature vector x, perform L2 regularization on it, and scale the feature vector to a unit vector in a high-dimensional spherical space; for each projection matrix Ri, right The matrix is calculated and the result vector is obtained, and the index corresponding to the largest value in the result vector is taken as the hash value hi of the feature vector; all the calculated matrix hash values hi are combined to obtain a set of hash values. It is used as a hash of image feature quantities, and images with the same hash value are divided into one category. Through this scheme, the image features can be extracted effectively, the image data set can be accurately divided, and the probability of hash collision can be reduced.
Description
technical field [0001] The invention relates to the field of deep learning image data set classification and processing, in particular to a classification and division method of image data sets. Background technique [0002] With the development of multimedia technology, image classification has become the focus of research in the field of computer vision. Image classification is to classify images into different preset categories according to certain attributes of images. How to effectively express images is a matter of improving The key to the accuracy of image classification, the selection and extraction of features is the current difficult problem in image classification. With the rapid development of mobile Internet, human society has entered the era of big data. Although traditional feature learning such as SIFT and HOG can extract some features of images and achieve good results in image classification, this artificial design feature method has certain defects. 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
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/55G06V10/764G06V10/40G06V10/82G06F17/16G06N3/04
CPCG06F16/55G06F17/16G06V10/40G06N3/045
Inventor 邓嘉新王亚强曹亮
Owner 成都成信高科信息技术有限公司