Early diagnosis of gastric cancer is crucial for improving patient survival rates. However, the limitations of traditional endoscopy in specificity and sensitivity contribute to the risks of misdiagnosis and missed diagnoses. The development of artificial intelligence (AI) technology based on convolutional neural networks (CNN) offers new approaches to enhancing the accuracy of endoscopic examinations. This paper systematically reviews the principles of CNNs and the application and challenges of computer-aided detection (CADe), computer-aided diagnosis (CADx), and computer-aided quality assessment (CADq) systems in the detection, diagnosis, and quality evaluation of early gastric cancer lesions. This lays the foundation for the widespread use of AI-assisted endoscopy in the early diagnosis of gastric cancer, with significant implications for reducing missed and misdiagnosed cases in clinical practice