Training method and device for image processing model and machine learning model

Abstract

According to the image processing model and machine learning model training method and device disclosed by the invention, the corresponding sample relationship identifier can be set for the target training sample according to the similarity between the sample characteristics of the first class of training samples in the first training sample set and the sample characteristics of at least part of training samples in the second training sample set; and training the model at least by using the target training sample with the sample relationship identifier. According to the invention, the number of training samples when the first training sample set is used for training the model can be increased, the training effect of the model is improved, and the sample characteristics of the training samples from the second training sample set can be learned when the model learns the sample characteristics of the training samples in the first training sample set. Therefore, the model can have universality in the scene corresponding to the first training sample set and the scene corresponding to the second training sample set.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to an image processing model and a training method and device for a machine learning model. Background technique [0002] With the development of artificial intelligence science and technology, related technologies of machine learning are constantly improving. The image recognition model obtained through machine learning can effectively identify the category of the target image. [0003] At present, if the image recognition model is to have high image recognition accuracy in multiple scenes, that is, it has versatility, the existing technology needs to collect training samples in multiple scenes and merge the collected training samples into In the same training sample set, the image recognition model is trained using the combined training sample set. For example, the image recognition model used in the enterprise to recognize the face of employees to check in to and from ...

AI Technical Summary

Problems solved by technology

[0004] However, the identification efficiency of artificial identification is low, and it is difficult to complete the identification of a large number of training samples, which may lead to insufficient samples of one or more categories of training samples in the combined training data set, making the image recognition model for the corresponding category The learning effect of the training samples is not enough, which leads to the insufficient recognition ability of the image recognition model for the corresponding category of images in different scenes, and cannot be universal.

Images