The paradigms of the generative adversarial networks (GANs), variational autoencoders (VAEs), transformer-based designs, and diffusion models, with the last one representing the state of the art in image generation models are discussed.
Abstract
Text-to-image generation is an increasingly fast-paced field of generative artificial intelligence, consisting of synthesizing images of high quality and semantic consistency based on natural language descriptions. In this paper, we give an extensive overview of the approach to text-to-image generation using deep learning, including the most common core model families, architecture designs, training approaches, and evaluation systems. We discuss the paradigms of the generative adversarial networks (GANs), variational autoencoders (VAEs), transformer-based designs, and diffusion models, with the last one representing the state of the art in image generation models. The review also discusses key aspects of pipelines such as text encoding, cross-modal alignment, mechanisms of attention, and decoding images. Popular datasets, methods, and metrics of evaluation, including Fréchet Inception Distance (FID) and CLIP-based similarity, are discussed. The application domains that involve creative content creation, medical imaging, education and industrial design are critically discussed. Despite significant advances, various issues still exist, such as low stability in training, excessive computational complexity, amplification of bias, generated images, and text–image alignment errors. Moral and social issues, such as misinformation, intellectual property, and equity, are critically examined. Lastly, we present future research directions to more controllable, more efficient and more interpretable text-to-image systems, focusing on multimodal foundation models and human–AI collaborative design.
A structured analytical perspective on the evolution of GAN-based T2I synthesis is provided, identifies key design trade-offs, and outlines open challenges for future research.
Shreedatta S. Sawant, S. Kaliraj, S. Raghavendra et al.· Discover Artificial Intellig...· 0 citations
: The core task of image generation models is to generate visual content that meets specific requirements based on given inputs. The wide application of artificial intelligence has accelerated the development of generative technologies, and image generation has had a significant impact across various fields in real-world scenarios. However, with many generative models available, selecting the appropriate one becomes a key issue. Current research mostly focuses on optimizing single models and lacks a comprehensive, systematic comparative analysis of multiple models. To address this, this paper conducts a multidimensional, standardized comparison of three generative models: Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs) and Diffusion Models, based on extensive literature review, combining theoretical analysis and experiments. The aim is to provide a scientific basis for model selection and support decision optimization in practical applications. Experimental results show that diffusion models produce better image quality but require higher hardware specifications. GANs and VAEs generate images of moderate quality with lower hardware requirements.
Qirui Guo, Yuxing Hu, Zhe-Jia Wang· Proceedings of the 3rd Inter...· 0 citations
An A2I model, AudioCanvas, fine-tuned on the A2I-Set is proposed, a unified, high-quality tri-modal dataset specifically designed for audio-visual research, including audio-conditioned image generation.
Dongxu Ge, Shansong Liu, Cheng Gong et al.· 0 citations
The proposed Attention-Based Deep Learning Pipeline of AI-Created Image Recognition incorporates three integrated branches, including low-level statistical feature extraction, high-level semantic representation learning, and attention-based feature refinement mechanism, which support the robustness and generalization ability of the proposed model in detecting AI-generated images in a variety of generators and conditions.
Nadia Ali· Al-Noor Journal of Engineeri...· 0 citations
An in-depth survey of fifteen state-of-art methodologies including classical CNN models, temporal-spatial video recognition, transformer-based networks, explainable AI (XAI) models, and models that combine multimodal large language model (LLM) products are provided.
Shavnam Shavnam, Neha Dhiman· International Journal of Inn...· 0 citations
In recent years, text-to-image (T2I) generation models have made substantial progress, particularly in visual realism and the expression of prompt semantics. However, a key difficulty remains: how to evaluate generated results automatically in a way that is both comprehensive and interpretable, while still being practical for real deployment. To address this issue, this paper proposes a multi-dimensional image quality assessment framework for T2I tasks. The framework examines generated images from five dimensions—text fidelity, perceptual quality, object consistency, relational consistency, and global semantic alignment—and derives a final quality score through normalization and weighted fusion. In terms of methodology, the framework combines Tesseract OCR, perceptual quality analysis based on Laplacian variance and exposure statistics, YOLO object detection, BLIP-based visual question answering, and CLIP image-text similarity, thereby forming a modular evaluation pipeline with diagnostic capability. Experiments on multiple mainstream T2I models and representative prompts show that the proposed method can not only distinguish overall performance differences across models, but also provide interpretable results at the level of individual dimensions.
