Rethinking Temporal Fusion with a Unified Gradient Descent View for 3D Semantic Occupancy Prediction

• URL: http://arxiv.org/abs/2504.12959v2
• Date: Fri, 18 Apr 2025 15:58:20 GMT
• Title: Rethinking Temporal Fusion with a Unified Gradient Descent View for 3D Semantic Occupancy Prediction
• Authors: Dubing Chen, Huan Zheng, Jin Fang, Xingping Dong, Xianfei Li, Wenlong Liao, Tao He, Pai Peng, Jianbing Shen,
• Abstract summary: We present GDFusion, a temporal fusion method for vision-based 3D semantic occupancy prediction (VisionOcc)<n>It opens up the underexplored aspects of temporal fusion within the VisionOcc framework, focusing on both temporal cues and fusion strategies.
• Score: 62.69089767730514
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We present GDFusion, a temporal fusion method for vision-based 3D semantic occupancy prediction (VisionOcc). GDFusion opens up the underexplored aspects of temporal fusion within the VisionOcc framework, focusing on both temporal cues and fusion strategies. It systematically examines the entire VisionOcc pipeline, identifying three fundamental yet previously overlooked temporal cues: scene-level consistency, motion calibration, and geometric complementation. These cues capture diverse facets of temporal evolution and make distinct contributions across various modules in the VisionOcc framework. To effectively fuse temporal signals across heterogeneous representations, we propose a novel fusion strategy by reinterpreting the formulation of vanilla RNNs. This reinterpretation leverages gradient descent on features to unify the integration of diverse temporal information, seamlessly embedding the proposed temporal cues into the network. Extensive experiments on nuScenes demonstrate that GDFusion significantly outperforms established baselines. Notably, on Occ3D benchmark, it achieves 1.4\%-4.8\% mIoU improvements and reduces memory consumption by 27\%-72\%.

Related papers

• SSRFlow: Semantic-aware Fusion with Spatial Temporal Re-embedding for Real-world Scene Flow [6.995663556921384]
  Scene flow provides the 3D motion field of the first frame from two consecutive point clouds. We propose a novel approach called Dual Cross Attentive (DCA) for the latent fusion and alignment between two frames based semantic contexts. We leverage novel domain adaptive losses to effectively bridge the gap of motion inference from synthetic to real-world.
  arXiv  Detail & Related papers  (2024-07-31T02:28:40Z)
• S^2Former-OR: Single-Stage Bi-Modal Transformer for Scene Graph Generation in OR [50.435592120607815]
  Scene graph generation (SGG) of surgical procedures is crucial in enhancing holistically cognitive intelligence in the operating room (OR) Previous works have primarily relied on multi-stage learning, where the generated semantic scene graphs depend on intermediate processes with pose estimation and object detection. In this study, we introduce a novel single-stage bi-modal transformer framework for SGG in the OR, termed S2Former-OR.
  arXiv  Detail & Related papers  (2024-02-22T11:40:49Z)
• Intensity Profile Projection: A Framework for Continuous-Time Representation Learning for Dynamic Networks [50.2033914945157]
  We present a representation learning framework, Intensity Profile Projection, for continuous-time dynamic network data. The framework consists of three stages: estimating pairwise intensity functions, learning a projection which minimises a notion of intensity reconstruction error. Moreoever, we develop estimation theory providing tight control on the error of any estimated trajectory, indicating that the representations could even be used in quite noise-sensitive follow-on analyses.
  arXiv  Detail & Related papers  (2023-06-09T15:38:25Z)
• Local-Global Temporal Difference Learning for Satellite Video Super-Resolution [55.69322525367221]
  We propose to exploit the well-defined temporal difference for efficient and effective temporal compensation. To fully utilize the local and global temporal information within frames, we systematically modeled the short-term and long-term temporal discrepancies. Rigorous objective and subjective evaluations conducted across five mainstream video satellites demonstrate that our method performs favorably against state-of-the-art approaches.
  arXiv  Detail & Related papers  (2023-04-10T07:04:40Z)
• Spatio-Temporal Self-Attention Network for Video Saliency Prediction [13.873682190242365]
  3D convolutional neural networks have achieved promising results for video tasks in computer vision. We propose a novel Spatio-Temporal Self-Temporal Self-Attention 3 Network (STSANet) for video saliency prediction.
  arXiv  Detail & Related papers  (2021-08-24T12:52:47Z)
• Searching Multi-Rate and Multi-Modal Temporal Enhanced Networks for Gesture Recognition [89.0152015268929]
  We propose the first neural architecture search (NAS)-based method for RGB-D gesture recognition. The proposed method includes two key components: 1) enhanced temporal representation via the 3D Central Difference Convolution (3D-CDC) family, and optimized backbones for multi-modal-rate branches and lateral connections. The resultant multi-rate network provides a new perspective to understand the relationship between RGB and depth modalities and their temporal dynamics.
  arXiv  Detail & Related papers  (2020-08-21T10:45:09Z)
• Spatiotemporal Fusion in 3D CNNs: A Probabilistic View [129.84064609199663]
  We propose to convert success thetemporal fusion strategies into a probability, which allows us to perform network-level evaluations of various fusion strategies without having to train them separately. Our approach greatly boosts the efficiency of analyzingtemporal fusion. We generate new fusion strategies which achieve the state-of-the-art performance on four well-grained action recognition datasets.
  arXiv  Detail & Related papers  (2020-04-10T10:40:35Z)
• A Graph Attention Spatio-temporal Convolutional Network for 3D Human Pose Estimation in Video [7.647599484103065]
  We improve the learning of constraints in human skeleton by modeling local global spatial information via attention mechanisms. Our approach effectively mitigates depth ambiguity and self-occlusion, generalizes to half upper body estimation, and achieves competitive performance on 2D-to-3D video pose estimation.
  arXiv  Detail & Related papers  (2020-03-11T14:54:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.