AI character generation method, device and equipment based on diffusion model

The invention discloses an AI character generation method and device based on a diffusion model, equipment and a storage medium, and the method comprises the steps: obtaining text image data of image-text pairing, carrying out the preprocessing of each piece of text image data, and obtaining a character image training data set; based on a preset loss function, inputting the character image training data set into a model constructed by a diffusion model, a Clip image encoder, a Clip text encoder and a character segmentation network for training to obtain a trained image-text generation model; and inputting a to-be-processed image and the text description words into the image-text generation model to obtain a result generation image containing text keywords. Through the scheme, the text content can be more stably and accurately generated in the image, the text content can be better embedded into a semantic scene, and efficient and automatic text image generation can be realized. 本发明公开了一种基于扩散模型的AI文字生成方法、装置、设备及存储