Optimal design models for visual communication: Image generation based on variational autoencoders


Abstract

Amidst the rapid evolution of modern visual design and human-computer interaction, the automatic synthesis of high-fidelity images from natural language descriptions has emerged as a pivotal research frontier in artificial intelligence within the domain of visual communication. Conventional text-to-image generation techniques continue to grapple with limitations in semantic-visual alignment, image fidelity, and model controllability, rendering them insufficient for complex, multi-scene, and multi-semantic generation tasks. In this study, we introduce VCL-GAN (Variational-Contrastive Learning GAN), a novel image generation framework that synergistically integrates a variational autoencoder with a contrastive learning paradigm to address the core challenges of language-driven image synthesis in visual communication optimization. Built upon the GAN architecture, the proposed model employs a dual-path semantic encoding strategy within the generator, wherein text representations derived via BERT are simultaneously infused into both a canonical semantic path and a structured latent path facilitated by the VAE. To further enrich semantic granularity, we incorporate cross-modal and intra-modal contrastive losses, thereby enhancing semantic coherence and perceptual fidelity in the generated outputs. Extensive experiments conducted on two benchmark datasets, CUB and MS-COCO, demonstrate that VCL-GAN surpasses state-of-the-art models such as AttnGAN in terms of Inception Score (IS) and Fréchet Inception Distance (FID), while also exhibiting superior robustness and generalizability in subjective user evaluations and ablation studies. The proposed approach not only advances the precision, clarity, and semantic richness of language-conditioned image generation but also presents a compelling paradigm for intelligent visual communication systems, holding significant promise for applications in AI-assisted design, cultural content creation, and multimodal human-computer interaction.
Ask to review this manuscript

Notes for potential reviewers

  • Volunteering is not a guarantee that you will be asked to review. There are many reasons: reviewers must be qualified, there should be no conflicts of interest, a minimum of two reviewers have already accepted an invitation, etc.
  • This is NOT OPEN peer review. The review is single-blind, and all recommendations are sent privately to the Academic Editor handling the manuscript. All reviews are published and reviewers can choose to sign their reviews.
  • What happens after volunteering? It may be a few days before you receive an invitation to review with further instructions. You will need to accept the invitation to then become an official referee for the manuscript. If you do not receive an invitation it is for one of many possible reasons as noted above.

  • PeerJ Computer Science does not judge submissions based on subjective measures such as novelty, impact or degree of advance. Effectively, reviewers are asked to comment on whether or not the submission is scientifically and technically sound and therefore deserves to join the scientific literature. Our Peer Review criteria can be found on the "Editorial Criteria" page - reviewers are specifically asked to comment on 3 broad areas: "Basic Reporting", "Experimental Design" and "Validity of the Findings".
  • Reviewers are expected to comment in a timely, professional, and constructive manner.
  • Until the article is published, reviewers must regard all information relating to the submission as strictly confidential.
  • When submitting a review, reviewers are given the option to "sign" their review (i.e. to associate their name with their comments). Otherwise, all review comments remain anonymous.
  • All reviews of published articles are published. This includes manuscript files, peer review comments, author rebuttals and revised materials.
  • Each time a decision is made by the Academic Editor, each reviewer will receive a copy of the Decision Letter (which will include the comments of all reviewers).

If you have any questions about submitting your review, please email us at [email protected].