Method for improving transportability of image recognition model

Abstract

The invention provides a method for improving transportability of an image recognition model. An improved minimum weight random search algorithm is used; endowing each attribute with a searched weightvalue; wherein the more the searched times are; weight increase, conversely, the smaller is, calculating the next searched probability of each attribute according to the weight; wherein the smaller the weight value is, the higher the searched probability is, otherwise, the lower the searched probability is, and further, according to the searched probability, the searching direction can be deviated to the attribute with the smaller weight value, namely, the attribute with the smaller searching frequency and the object with the larger weight are properly ignored, so that the purpose of searching balance is achieved; through an E-S judgment method, complex operation of further calculation accuracy is reduced, Meanwhile, the purpose of screening objects is achieved; by increasing the complexity of each attribute combination and applying a convolutional neural network which is constructed based on a Leaky Relu activation function and is provided with three convolutional layers, the purposeof 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 only with the human eye, but with the help of computer technology, through various processing and analysis of the image through the computer, and finally identifying the target we want to study. However, in the process of recognition, the computer cannot make an autonomous and dialectical judgment on the image like a human, and can only mechanically search for the characteristics of the image to complete the image recognition. Therefore, it is particularly important for the field of image recognition to train a perfect image recognition model. [0003] In research and experiment...

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

Images