Method and device for training image video quality evaluation model based on active learning

Abstract

The invention provides an image video quality evaluation model training method and device based on active learning. The method is suitable for the technical field of machine learning and the like, and comprises the following steps: determining the uncertainty of a certain to-be-evaluated sample by using a quality evaluation model based on an active learning mode under known k labeled samples (0<k); and based on the uncertainty of the to-be-evaluated sample, if the uncertainty of the sample is very high, preferentially performing subjective scoring on the sample, adding the sample into K marked samples, and retraining the quality evaluation model to achieve better performance. According to the method, active learning and an image video quality evaluation model are combined, more valuable samples are selected from a massive image video library to be provided for marking personnel to score through an active learning mode, on the premise that the model performances are the same, the number of samples needing subjective scoring can be reduced, and thus the marking cost is saved.

Description

technical field [0001] Embodiments of the present application relate to technical fields such as computer vision (image) or machine learning of artificial intelligence, and more specifically, to a training method and device for an image and video quality evaluation model based on active learning. Background technique [0002] The quality of the image / video can generally use the algorithm model to calculate the quality index of the video / image. [0003] Usually, by randomly selecting samples, active scoring is performed based on the selected samples, and then the quality evaluation model is trained based on the actively scored samples. However, for this method of training quality evaluation models, it is easy to select many worthless samples by randomly selecting samples, especially there will be a lot of similar and redundant data in the massive image and video database. Moreover, the number of samples to be selected needs to be set in advance, which is not easy to control....

AI Technical Summary

Problems solved by technology

However, for this method of training the quality evaluation model, the method of randomly selecting samples is easy to select many worthless samples, especially in the massive image and video database, there will be a lot of similar and redundant data

Moreover, the number of selected samples needs to be set in advance, which is not easy to control

In addition, completely isolating subjective scoring and model training will lead to the need for re-scoring if the quality of the dataset after subjective scoring is found to be low, which is time-consuming and labor-intensive with a low error tolerance rate

Images