ENHANCING IMAGE ACCURACY THROUGH ARTIFICIAL INTELLIGENCE: A COMPARATIVE STUDY OF CNN, GAN, AND TRANSFORMER MODELS

The demand for high-quality and precise image interpretation continues to grow across a wide range of domains, from medical diagnostics and security surveillance to remote sensing and autonomous systems. Conventional image processing methods, while effective in controlled conditions, often fall short when confronted with noise, complex textures, or variations in scale and illumination. In recent years, advances in artificial intelligence have opened new possibilities for overcoming these limitations by offering adaptive and data-driven approaches. This paper examines how modern learning-based techniques, including convolutional networks, transformer-based models, and generative frameworks, contribute to the improvement of image accuracy. Emphasis is placed on the integration of these methods with established preprocessing pipelines, as well as their comparative strengths in feature representation and enhancement tasks. Experimental evaluations conducted on benchmark datasets demonstrate consistent improvements in image fidelity and robustness compared with traditional baselines. The findings suggest that leveraging artificial intelligence not only enhances accuracy but also supports more generalizable and efficient solutions for future image processing applications.