Strictly-ID-Preserved and Controllable Accessory Advertising Image Generation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Xue, Youze, Chen, Binghui, Geng, Yifeng, Xie, Xuansong, Chen, Jiansheng, Ma, Hongbing
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913304705761280
author Xue, Youze
Chen, Binghui
Geng, Yifeng
Xie, Xuansong
Chen, Jiansheng
Ma, Hongbing
author_facet Xue, Youze
Chen, Binghui
Geng, Yifeng
Xie, Xuansong
Chen, Jiansheng
Ma, Hongbing
contents Customized generative text-to-image models have the ability to produce images that closely resemble a given subject. However, in the context of generating advertising images for e-commerce scenarios, it is crucial that the generated subject's identity aligns perfectly with the product being advertised. In order to address the need for strictly-ID preserved advertising image generation, we have developed a Control-Net based customized image generation pipeline and have taken earring model advertising as an example. Our approach facilitates a seamless interaction between the earrings and the model's face, while ensuring that the identity of the earrings remains intact. Furthermore, to achieve a diverse and controllable display, we have proposed a multi-branch cross-attention architecture, which allows for control over the scale, pose, and appearance of the model, going beyond the limitations of text prompts. Our method manages to achieve fine-grained control of the generated model's face, resulting in controllable and captivating advertising effects.
format Preprint
id arxiv_https___arxiv_org_abs_2404_04828
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Strictly-ID-Preserved and Controllable Accessory Advertising Image Generation
Xue, Youze
Chen, Binghui
Geng, Yifeng
Xie, Xuansong
Chen, Jiansheng
Ma, Hongbing
Computer Vision and Pattern Recognition
Customized generative text-to-image models have the ability to produce images that closely resemble a given subject. However, in the context of generating advertising images for e-commerce scenarios, it is crucial that the generated subject's identity aligns perfectly with the product being advertised. In order to address the need for strictly-ID preserved advertising image generation, we have developed a Control-Net based customized image generation pipeline and have taken earring model advertising as an example. Our approach facilitates a seamless interaction between the earrings and the model's face, while ensuring that the identity of the earrings remains intact. Furthermore, to achieve a diverse and controllable display, we have proposed a multi-branch cross-attention architecture, which allows for control over the scale, pose, and appearance of the model, going beyond the limitations of text prompts. Our method manages to achieve fine-grained control of the generated model's face, resulting in controllable and captivating advertising effects.
title Strictly-ID-Preserved and Controllable Accessory Advertising Image Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2404.04828