GTAD: Global Temporal Aggregation Denoising Learning for 3D Semantic Occupancy Prediction
This work addresses a key limitation in dynamic environment perception for autonomous driving systems, offering a novel approach to improve scene understanding.
The paper tackles the problem of inadequate temporal information utilization in 3D semantic occupancy prediction for autonomous driving by proposing GTAD, a global temporal aggregation denoising network that aggregates local and global temporal features, achieving superior performance on benchmarks like nuScenes and Occ3D-nuScenes.
Accurately perceiving dynamic environments is a fundamental task for autonomous driving and robotic systems. Existing methods inadequately utilize temporal information, relying mainly on local temporal interactions between adjacent frames and failing to leverage global sequence information effectively. To address this limitation, we investigate how to effectively aggregate global temporal features from temporal sequences, aiming to achieve occupancy representations that efficiently utilize global temporal information from historical observations. For this purpose, we propose a global temporal aggregation denoising network named GTAD, introducing a global temporal information aggregation framework as a new paradigm for holistic 3D scene understanding. Our method employs an in-model latent denoising network to aggregate local temporal features from the current moment and global temporal features from historical sequences. This approach enables the effective perception of both fine-grained temporal information from adjacent frames and global temporal patterns from historical observations. As a result, it provides a more coherent and comprehensive understanding of the environment. Extensive experiments on the nuScenes and Occ3D-nuScenes benchmark and ablation studies demonstrate the superiority of our method.