A method and system for intelligent recognition of target text in images
A target text and intelligent recognition technology, applied in character recognition, neural learning methods, character and pattern recognition, etc., can solve problems such as inability to accurately locate users, improve labeling accuracy, suppress invalid features, and reduce computational complexity Effect
Active Publication Date: 2022-08-05
烟台中科网络技术研究所 +1
View PDF9 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
This method can realize the automatic labeling of the text in the image, but if Figure 4 As shown, it can only mark the text area in the image in units of "lines", and cannot accurately locate the position of a single or several characters in the image that the user is interested in
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
[0057] The following combined with the attached drawings and the specific embodiments to further explain the present invention in detail;
[0058] like figure 1 As well as 2 Show, the target text intelligent identification method in the image includes the following steps:
[0059] Step S1: Get the image of the target text, express the target text from the pixel point coordinates, build a pixel coordinate recognition neural network, and train the pixel coordinate recognition neural network by entering the pixel coordinate values of all text in the target text to obtain text self -coding coding The characteristics of model and target text.
[0060] Specifically, (1) The user gives the target text of the target text based on personal needs, and uses the vertical line of the vertical line at the leftmost pixel in each text as the Y axis, and the horizontal line at the top of the pixel is the X -axis. Create a text pixel coordinate system based on the pixel distribution of each text ...
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 relates to the technical field of artificial intelligence, in particular to an intelligent recognition method for target text in an image. The steps of the method are as follows: represent the target text by pixel coordinates, input the pixel coordinates of the target text to train the pixel coordinate recognition neural network, and obtain the text self-encoding model and the expression features of the target text; calculate the background area according to the height value of the area number, extract the background area covered, and the remaining area is the foreground area; use the trained text auto-encoding model to obtain the expression characteristics of the text to be recognized in the foreground area, and compare the expression characteristics of the text with the expression characteristics of the expected target text. , if the error between the two reaches the preset threshold, the recognized text is the target text. The invention also provides an intelligent recognition system for target text in an image. The invention recognizes the target text based on the text self-encoding model, can accurately locate the position of the target text in the image, has low computational complexity and high recognition accuracy.
Description
Technical field [0001] The invention involves the field of artificial intelligence technology, and specially involves a target text intelligent identification method and system in the image. Background technique [0002] With the popularity of the Internet and smart terminals, the Internet applications based on social networks have developed rapidly, becoming the core carrier of mass information exchange and data release. As social data explosion increases, data acquisition and information analysis on the basis of this basis It has also become a research hotspot in the industrial and academic circles. [0003] The general thinking of the existing technical image recognition is to first position the text area in the image and perform OCR recognition of the text area content, and then match the text content to be labeled with the identified text content. Finally Automatically marked in the original image. For example, the patent number is Shen CN111310693A patent proposed a smart l...
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): G06V30/148G06V30/19G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V30/153G06V30/10G06N3/048G06N3/045G06F18/22
Inventor 段东圣王海洋佟玲玲时磊井雅琪段运强任博雅李真张旋宋吉锋王丽萍
Owner 烟台中科网络技术研究所