Integrating attention mechanisms in multi-scale image detection: a bibliometric analysis of research evolution and frontier trends

In recent years, image detection technology has faced many challenges, such as balancing efficiency and accuracy when dealing with complex environments, multi-scale, forgery generation, and dynamic scenes. In addition, the relevant research results in the field of deep learning are becoming increasingly diverse. Therefore, systematic bibliometric analysis is particularly necessary to fully understand the application progress and development trend of attention mechanisms in multi-scale image detection. This study uses bibliometric analysis methods to systematically analyze the historical evolution and current research mode of attention mechanism in multi-scale image detection and predict future development trends. Specifically, this study evaluates 506 academic journal articles from the Web of Science database, identifies influential publications, constructs research patterns, and predicts emerging trends through co-citation and co-word analysis. This study identifies four different clusters for co-citation analysis and four different clusters for co-word analysis. Although the attention mechanism shows great potential in modelling complex cross-modal associations, especially in image detection tasks, it still faces challenges such as gradient diffusion and computational efficiency in multi-scale and dynamic scene adaptability. The analysis of this study provides a valuable reference for the development of related technologies in the future.