Point Cloud Saliency

The rapid advancements in immersive multimedia technologies have significantly reshaped how users interact with and experience digital content. Central to this transformation is the integration of 3D point clouds—a versatile representation that captures the intricate geometry of real-world environments with exceptional precision. These point clouds have found widespread application in fields such as Virtual Reality (VR), Augmented Reality (AR), autonomous driving, and cultural heritage preservation. However, despite their detailed representation capabilities, point clouds pose considerable challenges in terms of storage, transmission, and computational processing due to their large volume and complexity. A key aspect of enhancing user interaction with point clouds lies in understanding visual attention mechanisms within immersive environments. Human perception naturally prioritizes specific regions of a scene, and this selective focus plays a crucial role in improving the Quality of Experience (QoE). Traditional approaches for analyzing visual attention have primarily been developed for 2D content, and their extension to 3D point cloud data remains a relatively underexplored domain. To address this gap, a novel framework is proposed that adapts existing 2D saliency models to the spatial characteristics of 3D data using orthographic projections. By bridging the methodological divide between 2D and 3D saliency prediction, this approach aims to facilitate a more accurate analysis of user attention in immersive 3D environments. The framework is evaluated using a 3D point cloud eye-tracking dataset. Various projection configurations and thresholding strategies are examined to ensure the effective transfer of saliency information from 2D to 3D representations. The results indicate that embedding visual saliency models into 3D point clouds can significantly enhance QoE by allocating computational resources to perceptually significant areas. This optimization contributes to lower latency and a more immersive user experience. Overall, the findings advance the field of visual saliency by offering new perspectives on improving the interaction between users and 3D content in immersive multimedia systems.