Image processing model training method, image generation method, device and equipment

Abstract

The embodiment of the invention relates to an image processing model training method, an image generation method, a device and equipment. The training method comprises the following steps: determining a first adjustable attribute feature vector and a first fixed attribute feature vector of a target object on a sample image; performing vector fusion on the first adjustable attribute feature vector and the first fixed attribute feature vector, inputting the fused vectors into a generator in a generative adversarial network, and generating a target image; determining a second adjustable attribute feature vector and a second fixed attribute feature vector of the target object on the target image; determining an adjustable attribute difference degree based on the first adjustable attribute feature vector and the second adjustable attribute feature vector; determining a fixed attribute difference degree based on the first fixed attribute feature vector and the second fixed attribute feature vector; and adjusting network parameters of the image processing model based on the adjustable attribute difference degree and the fixed attribute difference degree. According to the embodiment of the invention, the continuous processing of the attributes of the target object and the state control of the specific attributes of the target object are realized.

Description

technical field [0001] The present disclosure relates to the technical field of image processing, and in particular to an image processing model training method, image generation method, device and equipment. Background technique [0002] Currently, image generation techniques are widely studied as an efficient way to synthesize new images. Among them, manipulating the specific properties of object instances on the image is not only conducive to more controllable image editing, but also beneficial to many image understanding tasks, such as in the field of unmanned driving or automatic driving, for the object instance on the image. State Estimation and Identification. [0003] Manipulating specific properties of an object instance on an image can be viewed as a transition of properties from one state to another, and considering states as domains / types, property manipulation can be modeled as domain / type transitions. In the existing attribute conversion scheme based on model...

AI Technical Summary

Problems solved by technology

In the existing attribute conversion scheme based on model training, on the one hand, due to the limitation of the model volume and the number of training samples in each attribute state, for example, using sample images of object instances in some specific attribute states for model training, etc., trained The model can only handle a limited number of discrete states of object instance attributes; on the other hand, since the model can only adapt to the discrete states of object instance attributes, the model processing process cannot well integrate the attributes to be adjusted and other attribute information of object instances Separation, so that state control and image generation only for a specific attribute cannot be realized

Images