<![CDATA[The increasing complexity of solid waste requires intelligent and efficient systems to support waste detection, classification, and management. This study aims to synthesize the development of computer vision-based waste detection and classification technologies and examine their contribution to smart waste management systems. The study employed a structured literature-based design using 19 scientific publications published between 2018 and 2026. The selected publications were analyzed through qualitative content analysis and comparative synthesis based on technological orientation, computational approach, primary task, implementation context, and reported contribution. The findings indicate a technological progression from convolutional neural network-based image classification and transfer learning toward object localization, real-time detection, and YOLO-based approaches. Recent studies also demonstrate increasing integration between computer vision, Internet of Things infrastructure, embedded computing, automated sorting, and cloud-based deployment. The analysis identifies dataset diversity, environmental variability, object overlap, computational efficiency, and deployment architecture as recurring challenges affecting practical implementation. The findings suggest that computer vision should be developed as an integrated component of smart waste management systems rather than as an isolated classification model. Effective implementation requires alignment between visual recognition capability, computational resources, communication infrastructure, and operational requirements. This study provides a structured synthesis of technological developments and highlights directions for developing more adaptive, scalable, and practically applicable computer vision-based waste management systems.]]>
