Multistep instructions, such as recipes and how-to guides, greatly benefit from visual aids, such as a series of images that accompany the instruction steps. While Large Language Models (LLMs) have become adept at generating coherent textual steps, Large Vision/Language Models (LVLMs) are less capable of generating accompanying image sequences. The most challenging aspect is that each generated image needs to adhere to the relevant textual step instruction, as well as be visually consistent with earlier images in the sequence.
To address this problem, we propose an approach for generating consistent image sequences, which integrates a Latent Diffusion Model (LDM) with an LLM to transform the sequence into a caption to maintain the semantic coherence of the sequence. In addition, to maintain the visual coherence of the image sequence, we introduce a copy mechanism to initialise reverse diffusion processes with a latent vector iteration from a previously generated image from a relevant step. Both strategies will condition the reverse diffusion process on the sequence of instruction steps and tie the contents of the current image to previous instruction steps and corresponding images.
Experiments show that the proposed approach is preferred by humans in 46.6% of the cases against 26.6% for the second best method. In addition, automatic metrics showed that the proposed method maintains semantic coherence and visual consistency across the sequence of visual illustrations.
(click on an image to enlarge)
@misc{bordalo2024generating,
title={Generating Coherent Sequences of Visual Illustrations for Real-World Manual Tasks},
author={João Bordalo and Vasco Ramos and Rodrigo Valério and Diogo Glória-Silva and Yonatan Bitton and Michal Yarom and Idan Szpektor and Joao Magalhaes},
year={2024},
eprint={2405.10122},
archivePrefix={arXiv},
primaryClass={cs.CV}
}