The rapid advancement of deep learning, particularly Convolutional Neural Networks (CNNs), has revolutionized image recognition. However, the computational intensity and power demands of CNNs present significant challenges for real-time, embedded, and edge applications. Field Programmable Gate Arrays (FPGAs) have emerged as a promising hardware platform to accelerate CNN inference, offering a balance between performance, flexibility, and energy efficiency. This paper presents a comprehensive overview of FPGA-based CNN acceleration for image recognition, covering architectural design, implementation strategies, performance analysis, and future directions. Recent research and practical implementations are discussed, highlighting the benefits and challenges of deploying CNNs on FPGAs.