A Method to Improve the Portability of Image Recognition Models
An image recognition and image technology, applied in the field of computer vision, can solve the problems of recognition, low model portability, poor performance of image recognition models, etc., and achieve the effect of reducing complex operations and improving portability
Active Publication Date: 2021-03-23
BEIJING UNIV OF CIVIL ENG & ARCHITECTURE
View PDF7 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Although both data sets are used to train digital recognition models, there is a large difference between the image characteristics of the two data sets. The computer cannot recognize the two data sets dialectically like a human being, and cannot directly judge the corresponding As a result, the image recognition model that has been trained on a specific data set does not perform well on other data sets, that is, the portability of the model is too low
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
[0041] Embodiments of the present invention are described in detail below, examples of which are shown in the drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary only for explaining the present invention and should not be construed as limiting the present invention.
[0042] Those skilled in the art will understand that unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the word "comprising" used in the description of the present invention refers to the presence of said features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, Integers, steps, operations, elements, components, and / or groups thereof. It will be understoo...
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 method for improving the portability of an image recognition model. Through the improved minimum weight random search algorithm, each attribute is given a searched weight value. The more times it is searched, the greater the weight, and vice versa. Small, according to the weight to calculate the probability of each attribute being searched next, the smaller the weight value is, the greater the probability of being searched next is, otherwise the probability of being searched is smaller, further according to the probability of being searched, the search direction can be adjusted It is biased towards attributes with smaller weight values, that is, attributes that have been searched for less times, and appropriately "ignores" objects with larger weights to achieve the purpose of search balance; the E‑S judgment method reduces the complexity of further calculation accuracy At the same time, the purpose of screening objects is also achieved; by increasing the complexity of each attribute combination and using a convolutional neural network with 3 convolutional layers constructed based on the Leaky Relu activation function, the purpose of fully extracting image features is achieved.
Description
technical field [0001] The invention relates to the technical field of computer vision, in particular to a method for improving the portability of an image recognition model. Background technique [0002] Image recognition is an important field of artificial intelligence, and it has been widely used and developed in machine learning. However, the image recognition referred to in this article is not just using human eyes, but using computer technology to identify, and through the computer to make various processing and analysis of the image, and finally identify the target we want to study. However, in the process of recognition, computers cannot make independent dialectical judgments on images like humans. They can only mechanically find the characteristics of images to complete image recognition. Therefore, training a perfect image recognition model is particularly important in the field of image recognition. [0003] In research and experiments, it will be found that if ...
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): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/047G06N3/045G06F18/2415G06F18/214
Inventor 谭志刘兴业曹红玉
Owner BEIJING UNIV OF CIVIL ENG & ARCHITECTURE