Introduction to the Special Issue on Advances in AI-Driven Computational Modeling for Image Processing

Attention-Based Deep Learning Pipeline for AI-Created Image Recognition

The advent of generative AI models that can generate some of the most realistic images including DALL·E, Midjourney, and Stable Diffusion has posed a significant threat to digital forensics and media authenticity verifications. Conventional methods of forgery detection, which focus on splicing or copy-move manipulations, are not very efficient in the case of AI-generated images, which do not contain any explicit tampering traces. This paper presents Attention-Based Deep Learning Pipeline of AI-Created Image Recognition, intended to resolve the increasing challenge of separating AI-synthesized and authentic imagery. The proposed architecture incorporates three integrated branches, including low-level statistical feature extraction, high-level semantic representation learning, and attention-based feature refinement mechanism. The features that are extracted are combined and fed through a fully connected classifier to generate final predictions. The Synthbuster dataset that consists of images of several generative models, including DALL·E 2, DALL·E 3, Firefly, Glide, and versions of Stable Diffusion, were experimentally evaluated. The suggested framework score had an average AUC of 97.4%, as well as high accuracy, precision, recall, and F1-score. These findings support the robustness and generalization ability of the proposed model in detecting AI-generated images in a variety of generators and conditions.

From Convolution to Attention and Beyond: A Systematic Review of Modern Vision Architectures

A critical review of computer vision, illustrating how architectural design, learning paradigms, and evaluation practices have co-evolved over time to facilitate more flexible and scalable systems, and outlining new research directions.

Generative Artificial Intelligence for image synthesis using Generative Adversarial Networks (GANs) and variational autoencoders

In recent years, the rapid advancement of Generative Artificial Intelligence (GenAI) has transformed the landscape of digital content creation, enabling high-fidelity image synthesis across various fields including healthcare, art, and computer vision. However, the proliferation of synthesized images has introduced critical challenges, specifically the need to distinguish real physical imagery from synthetic, AI-generated counterfeits. This research paper presents a comprehensive, end-to-end framework that addresses both generative synthesis and discriminative detection under hardware-constrained (CPU-only) environments. We implement two synthesis methodologies like a Deep Convolutional Generative Adversarial Network (DCGAN) and a Convolutional Variational Autoencoder (ConvVAE) - trained on real image distributions to generate synthetic data. Concurrently, we present a compact Convolutional Neural Network (MiniCNNClassifier) designed to detect and classify images as real or fake. The framework is validated using a balanced dataset of 60,000 images (30,000 real and 30,000 synthetic). Our preprocessing pipeline ensures uniform size and resolution across heterogeneous inputs. Experimental results demonstrate that the MiniCNNClassifier achieves an outstanding validation accuracy of 98.7% and a Precision of 99.5%, Recall of 97.8%, F1-score of 98.6% in detecting fake samples. Furthermore, we provide a qualitative and quantitative comparison of DCGAN and ConvVAE architectures, discussing trade-offs between training stability and sample fidelity. Finally, we host the models on an interactive Streamlit-based web interface to enable real-time generation and classification.

PixelBoost 8 – Pixel Quality with 8X Highlights Boosting Enhancement

In recent years, deep learning has become a fundamental technology across a wide array of scientific and industrial fields, largely fuelled by advances in computational capabilities. One area that has experienced substantial progress is face hallucination—the task of improving the resolution of facial images. This process is critical to various computer vision applications, including facial recognition, feature extraction, and identity verification. Recently, deep generative models, particularly Generative Adversarial Networks (GANs), have led the field. Although these models have produced remarkable results, there is still a pressing need to further improve both accuracy and output quality. In order to address these problems, we propose a new GAN-based face hallucination method. This method is primarily based on the Enhanced Super-Resolution Generative Adversarial Network (ESRGAN). We present a personalised adaptation of ESRGAN that employs the VGG16 architecture with a compact pre-trained version. This method balances output image quality and computational efficiency. Experiments show that our approach is effective. The improved model obtains a maximum peak signal-to-noise ratio (PSNR) of 30.30. The Learned Perceptual Image Patch Similarity (LPIPS) score is 0.0817, whereas the Structural Similarity Index Measure (SSIM) is 0.8757. The results surpass many state-of-the-art methods available today. These enhancements have a significant impact and importance.

Text-to-Image Generation via Deep Learning: A Comprehensive Review of Models, Architectures, and Future Directions

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.

Detection of AI Generated Faces and Real Human Faces Using Deep Learning-Based Transfer Learning