Text Based Image Creator

This paper has presented PixelMind, a locally deployable, GPU-accelerated text-to-image web application built on Stable Diffusion v1.5. The system meets all ten of its defined objectives and achieves 100% test coverage across 18 test cases spanning nine functional modules. By running entirely on an NVIDIA GeForce GTX 1650, PixelMind demonstrates that production-quality generative AI tooling does not require managed cloud infrastructure, expensive API subscriptions, or specialised hardware beyond what is already present in many student and researcher workstations. The architecture is modular, privacy-preserving, and free to operate. Its FID score of 8.59 compares favourably with DALL·E 2's 10.39, while voice input, seven real-time image filters, a personal gallery, usage analytics, and a secured admin dashboard give PixelMind a feature depth that sets it apart from the bare-bones local inference scripts available in the open-source community. The project offers a replicable reference architecture for institutions, educators, and independent developers who want to democratise access to generative AI without compromising user privacy or incurring ongoing service costs.

[1]    I. Goodfellow et al., "Generative Adversarial Nets," Advances in Neural Information Processing Systems (NeurIPS), vol. 27, 2014. [2]   S. Reed et al., "Generative Adversarial Text to Image Synthesis," Proc. 33rd Int. Conf. Machine Learning (ICML), 2016. [3]    H. Zhang et al., "StackGAN: Text to Photo-realistic Image Synthesis with Stacked Generative Adversarial Networks," IEEE Int. Conf. Computer Vision (ICCV), 2017. [4]   T. Xu et al., "AttnGAN: Fine-Grained Text to Image Generation with Attentional Generative Adversarial Networks," IEEE/CVF CVPR, 2018. [5]      D. P. Kingma and M. Welling, "Auto-Encoding Variational Bayes," arXiv:1312.6114, 2013. [6]         A. Ramesh et al., "Zero-Shot Text-to-Image Generation," Proc. 38th Int. Conf. Machine Learning (ICML), 2021. [7]    J. Ho, A. Jain, and P. Abbeel, "Denoising Diffusion Probabilistic Models," Advances in Neural Information Processing Systems (NeurIPS), vol. 33, 2020. [8]  A. Ramesh et al., "Hierarchical Text-Conditional Image Generation with CLIP Latents," arXiv:2204.06125, OpenAI, 2022. [9]       C. Saharia et al., "Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding," arXiv:2205.11487, Google Research, 2022. [10]  R. Rombach et al., "High-Resolution Image Synthesis with Latent Diffusion Models," IEEE/CVF CVPR, 2022.