Modern computer vision from CNNs to foundation models: a unified framework for representation learning

Modern computer vision is increasingly driven by the search for representations that generalize across tasks, modalities, and learning settings, rather than excelling at a single benchmark. This has produced a broad collection of architectures and paradigms, including convolutional neural networks (CNNs), Vision Transformers (ViTs), foundation models, and diffusion-based generative models. Despite differing in design and training objectives, these approaches increasingly draw on a common set of principles: representation learning, self-attention, self-supervision, multimodal alignment, and generative modeling. As a result, the boundaries between discriminative, multimodal, and generative paradigms are becoming progressively less distinct. This survey examines that convergence through the lens of representation learning. We first introduce a common mathematical framework for describing vision architectures and learning paradigms across settings, then trace the evolution from CNNs and Vision Transformers, through hybrid CNN-Transformer architectures, foundation models, and multimodal vision-language systems, to recent generative approaches, including diffusion models, Diffusion Transformers, and Rectified Flows. By analyzing these paradigms within a single framework, we show how discriminative, multimodal, and generative learning are converging around shared representational principles, argue for a unified representation-learning perspective on modern computer vision, and outline open challenges and future directions toward general-purpose visual intelligence, multimodal foundation models, and trustworthy deployment.