Hierarchical Aggregation for 3D Instance Segmentation

Chen, Shaoyu; Fang, Jiemin; Zhang, Qian; Liu, Wenyu; Wang, Xinggang

doi:10.1109/iccv48922.2021.01518

Cited by 113 publications

(72 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When replacing colors with multiview features and normals (last row), our PointGroup implementation significantly outperforms the original one. Our multiview-based PointGroup gives mAP@0.5 of 62.8, which is close to the performance of the current state-of-the-art model HAIS[9], which achieves 64.1 on the validation set of ScanNet v2. Our implementation also surpasses the original PointGroup in training speed: given point coordinates and colors as input, it takes less than two days to train the model in our implementation, while the original one could take up to three days until convergence.…”

supporting

confidence: 64%

D3Net: A Unified Speaker-Listener Architecture for 3D Dense Captioning and Visual Grounding

Chen¹,

Wu²,

Nießner³

et al. 2021

Preprint

View full text Add to dashboard Cite

Listener discriminates" this is a black chair . it is against the wall and facing the door. " 4 " this chair is black . it is under the whiteboard facing the round table ."2 " … " 3 " … " 5 " … " 1 Dense captions Detections Figure 1. We introduce D 3 Net, an end-to-end neural speaker-listener architecture that can detect, describe and discriminate. D 3 Net also enables semi-supervised training on Scan-Net data with partially annotated descriptions.

show abstract

supporting

confidence: 64%

D3Net: A Unified Speaker-Listener Architecture for 3D Dense Captioning and Visual Grounding

Chen¹,

Wu²,

Nießner³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Jiang et al [14] propose Point-Group to segment objects on original and offset-shifted point sets, relying on a simple yet effective algorithm that groups nearby points of the same label and expands the group progressively. Chen et al [4] extend PointGroup and propose HAIS that further absorbs surrounding fragments of instances and then refines the instances based on intrainstance prediction. Liang et al [17] SSTNet to construct a tree network from pre-computed superpoints then traverse the tree and split nodes to get object instances.…”

Section: Related Workmentioning

confidence: 99%

“…A semantic branch is constructed from a two-layer MLP and learns to output semantic scores S = {s 1 , ..., s N } ∈ R N ×Nclass for N points over N class classes. Different from existing methods [4,14], we directly perform grouping on semantic scores without converting the semantic scores to one-hot semantic predictions.…”

Section: Point-wise Prediction Networkmentioning

confidence: 99%

“…where s * is the semantic label, o * is offset label representing the vector from a point to the geometric center of the instance that the point belongs to (analogous to [4,14,17]), and 1 {p i } is the indicator function indicating whether the point p i belongs to any instance.…”

Section: Point-wise Prediction Networkmentioning

confidence: 99%

“…State-of-the-art methods [4,14,17] consider 3D instance segmentation as a bottom-up pipeline. They learn the point-wise semantic labels and center offset vectors and then group points of the same labels with small geomet- ric distances into instances.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

SoftGroup for 3D Instance Segmentation on Point Clouds

Vu¹,

Kookhoi²,

Luu³

et al. 2022

Preprint

View full text Add to dashboard Cite

Existing state-of-the-art 3D instance segmentation methods perform semantic segmentation followed by grouping. The hard predictions are made when performing semantic segmentation such that each point is associated with a single class. However, the errors stemming from hard decision propagate into grouping that results in (1) low overlaps between the predicted instance with the ground truth and (2) substantial false positives. To address the aforementioned problems, this paper proposes a 3D instance segmentation method referred to as SoftGroup by performing bottom-up soft grouping followed by top-down refinement. SoftGroup allows each point to be associated with multiple classes to mitigate the problems stemming from semantic prediction errors and suppresses false positive instances by learning to categorize them as background. Experimental results on different datasets and multiple evaluation metrics demonstrate the efficacy of SoftGroup. Its performance surpasses the strongest prior method by a significant margin of +6.2% on the ScanNet v2 hidden test set and +6.8% on S3DIS Area 5 in terms of AP 50 . Soft-Group is also fast, running at 345ms per scan with a single Titan X on ScanNet v2 dataset. The source code and trained models for both datasets are available at https: //github.com/thangvubk/SoftGroup.git.

show abstract

3D Instances as 1D Kernels

Shi

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Hierarchical Aggregation for 3D Instance Segmentation

Cited by 113 publications

References 31 publications

D3Net: A Unified Speaker-Listener Architecture for 3D Dense Captioning and Visual Grounding

D3Net: A Unified Speaker-Listener Architecture for 3D Dense Captioning and Visual Grounding

SoftGroup for 3D Instance Segmentation on Point Clouds

3D Instances as 1D Kernels

Contact Info

Product

Resources

About