TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Papers/PanopticNDT: Efficient and Robust Panoptic Mapping

PanopticNDT: Efficient and Robust Panoptic Mapping

Daniel Seichter, Benedict Stephan, Söhnke Benedikt Fischedick, Steffen Müller, Leonard Rabes, Horst-Michael Gross

2023-09-24Panoptic Segmentation (PanopticNDT instances)Panoptic Segmentation2D Panoptic SegmentationScene Understanding3D Panoptic SegmentationScene Classification (unified classes)Semantic Segmentation3D Semantic Segmentation
PaperPDFCodeCodeCode(official)Code

Abstract

As the application scenarios of mobile robots are getting more complex and challenging, scene understanding becomes increasingly crucial. A mobile robot that is supposed to operate autonomously in indoor environments must have precise knowledge about what objects are present, where they are, what their spatial extent is, and how they can be reached; i.e., information about free space is also crucial. Panoptic mapping is a powerful instrument providing such information. However, building 3D panoptic maps with high spatial resolution is challenging on mobile robots, given their limited computing capabilities. In this paper, we propose PanopticNDT - an efficient and robust panoptic mapping approach based on occupancy normal distribution transform (NDT) mapping. We evaluate our approach on the publicly available datasets Hypersim and ScanNetV2. The results reveal that our approach can represent panoptic information at a higher level of detail than other state-of-the-art approaches while enabling real-time panoptic mapping on mobile robots. Finally, we prove the real-world applicability of PanopticNDT with qualitative results in a domestic application.

Results

TaskDatasetMetricValueModel
Semantic SegmentationScanNettest mIoU68.1PanopticNDT (10cm)
Semantic SegmentationScanNetval mIoU68.39PanopticNDT (10cm)
Semantic SegmentationNYU Depth v2Mean IoU59.02EMSANet (2x ResNet-34 NBt1D, PanopticNDT version, finetuned)
Semantic SegmentationHypersimmIoU49.74EMSANet (2x ResNet-34 NBt1D)
Semantic SegmentationHypersimmIoU (test)46.66EMSANet (2x ResNet-34 NBt1D)
Semantic SegmentationNYU Depth v2PQ51.15EMSANet (2x ResNet-34 NBt1D, PanopticNDT version, finetuned)
Semantic SegmentationScanNetV2PQ59.19PanopticNDT (10cm)
Semantic SegmentationHypersimPQ34.95EMSANet (2x ResNet-34 NBt1D)
Semantic SegmentationHypersimPQ (test)29.77EMSANet (2x ResNet-34 NBt1D)
Semantic SegmentationHypersimmIoU49.12EMSANet (2x ResNet-34 NBt1D)
Semantic SegmentationHypersimmIoU (test)44.66EMSANet (2x ResNet-34 NBt1D)
Semantic SegmentationHypersimmIoU45.43PanopticNDT (10cm)
Semantic SegmentationHypersimmIoU (test)45.34PanopticNDT (10cm)
Semantic SegmentationHypersimmIoU44.31SemanticNDT (10cm)
Semantic SegmentationHypersimmIoU (test)44.8SemanticNDT (10cm)
3D Semantic SegmentationHypersimmIoU45.43PanopticNDT (10cm)
3D Semantic SegmentationHypersimmIoU (test)45.34PanopticNDT (10cm)
3D Semantic SegmentationHypersimmIoU44.31SemanticNDT (10cm)
3D Semantic SegmentationHypersimmIoU (test)44.8SemanticNDT (10cm)
10-shot image generationScanNettest mIoU68.1PanopticNDT (10cm)
10-shot image generationScanNetval mIoU68.39PanopticNDT (10cm)
10-shot image generationNYU Depth v2Mean IoU59.02EMSANet (2x ResNet-34 NBt1D, PanopticNDT version, finetuned)
10-shot image generationHypersimmIoU49.74EMSANet (2x ResNet-34 NBt1D)
10-shot image generationHypersimmIoU (test)46.66EMSANet (2x ResNet-34 NBt1D)
10-shot image generationNYU Depth v2PQ51.15EMSANet (2x ResNet-34 NBt1D, PanopticNDT version, finetuned)
10-shot image generationScanNetV2PQ59.19PanopticNDT (10cm)
10-shot image generationHypersimPQ34.95EMSANet (2x ResNet-34 NBt1D)
10-shot image generationHypersimPQ (test)29.77EMSANet (2x ResNet-34 NBt1D)
10-shot image generationHypersimmIoU49.12EMSANet (2x ResNet-34 NBt1D)
10-shot image generationHypersimmIoU (test)44.66EMSANet (2x ResNet-34 NBt1D)
10-shot image generationHypersimmIoU45.43PanopticNDT (10cm)
10-shot image generationHypersimmIoU (test)45.34PanopticNDT (10cm)
10-shot image generationHypersimmIoU44.31SemanticNDT (10cm)
10-shot image generationHypersimmIoU (test)44.8SemanticNDT (10cm)
Panoptic SegmentationNYU Depth v2PQ51.15EMSANet (2x ResNet-34 NBt1D, PanopticNDT version, finetuned)
Panoptic SegmentationScanNetV2PQ59.19PanopticNDT (10cm)
Panoptic SegmentationHypersimPQ34.95EMSANet (2x ResNet-34 NBt1D)
Panoptic SegmentationHypersimPQ (test)29.77EMSANet (2x ResNet-34 NBt1D)
Panoptic SegmentationHypersimmIoU49.12EMSANet (2x ResNet-34 NBt1D)
Panoptic SegmentationHypersimmIoU (test)44.66EMSANet (2x ResNet-34 NBt1D)
2D Panoptic SegmentationScanNetV2PQ58.22EMSANet (2x ResNet-34 NBt1D, PanopticNDT version)

Related Papers

SeC: Advancing Complex Video Object Segmentation via Progressive Concept Construction2025-07-21Advancing Complex Wide-Area Scene Understanding with Hierarchical Coresets Selection2025-07-17Argus: Leveraging Multiview Images for Improved 3-D Scene Understanding With Large Language Models2025-07-17City-VLM: Towards Multidomain Perception Scene Understanding via Multimodal Incomplete Learning2025-07-17DiffOSeg: Omni Medical Image Segmentation via Multi-Expert Collaboration Diffusion Model2025-07-17SCORE: Scene Context Matters in Open-Vocabulary Remote Sensing Instance Segmentation2025-07-17Unified Medical Image Segmentation with State Space Modeling Snake2025-07-17A Privacy-Preserving Semantic-Segmentation Method Using Domain-Adaptation Technique2025-07-17