Convolutional Neural Networks (CNNs) have transformed the field of computer vision by enabling automated feature extraction and delivering remarkable performance across a wide range of visual recognition tasks. The rapid evolution of CNN architectures, coupled with the growing availability of large-scale datasets and high-performance computing platforms, has accelerated their adoption in domains such as image classification, object detection, semantic segmentation, medical image analysis, autonomous driving, precision agriculture, and intelligent surveillance. Despite these advances, the increasing complexity of network designs and the emergence of alternative deep learning paradigms have created the need for a unified and up-to-date assessment of CNN-based developments. This review presents a comprehensive analysis of CNNs for computer vision by systematically examining their architectural evolution, major application areas, and recent research trends. Relevant studies were identified through a structured literature search across leading scientific databases using predefined inclusion and exclusion criteria to ensure the selection of high-quality contributions. The review highlights the progression from early CNN models to modern efficient architectures, compares their strengths and limitations, and summarizes their practical impact across diverse application domains. Furthermore, key challenges, including computational cost, data dependency, model interpretability, robustness, and deployment constraints, are critically discussed. Emerging directions such as lightweight CNNs, hybrid CNN--Transformer frameworks, self-supervised learning, edge intelligence, and explainable artificial intelligence are also explored. By integrating architectural, application-oriented, and future perspectives, this review provides a consolidated reference for researchers and practitioners while identifying promising opportunities for the next generation of computer vision systems.

Convolutional Neural Networks, Computer Vision, Deep Learning, Object Detection, Image Classification, Transfer Learning, Vision Transformers, Medical Imaging