Salient Object Detection for Point Clouds

Fan, Songlin; Gao, Wei; Li, Ge

doi:10.1007/978-3-031-19815-1_1

Cited by 12 publications

(11 citation statements)

References 58 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2 shows the pipeline. For the specific structures of FAB, PPB, and SPB, please refer to (Fan, Gao, and Li 2022).…”

Section: Proposed Methodology Baseline Modelsmentioning

confidence: 99%

“…However, most of these works (Tian et al 2021) are limited to RGB images and ignore 3D spatial geometry of objects. Therefore, researchers extend SOD to 3D scenarios, including RGBD images (Fan et al 2020) and point clouds (Fan, Gao, and Li 2022). For RGBD images, (Zhang et al 2020b) adopts a two-stream architecture to process RGB features and depth features separately.…”

Section: Related Workmentioning

confidence: 99%

“…For point clouds, previous works (Zheng et al 2019) mainly aim to predict the distribution heatmap of human attention, failing to segment complete salient objects. Given this, Fan et al (Fan, Gao, and Li 2022) (Charles et al 2017) and PointNet++ (Qi et al 2017) propose to process raw points directly and achieve surprising performance across multiple tasks. Following PointNet, PointCNN (Li et al 2018) proposes to regularize the points with an X-transformation and then process them with typical convolutions.…”

Section: Related Workmentioning

confidence: 99%

“…Point Cloud Salient Object Detection (PCSOD) is a 3D segmentation task newly proposed by Fan et al (Fan, Gao, and Li 2022), which aims to segment the most attractive objects in point cloud scenarios. Compared with 2D salient object detection (Wei, Wang, and Huang 2020;Wei et al 2020b), PCSOD can make full use of depth information, thus reducing the prediction ambiguity.…”

Section: Introductionmentioning

confidence: 99%

“…Otherwise, this sample will be ignored to avoid saliency conflict. For more comparison between PCSOD and 3DSS, and data annotation details, please refer to (Fan, Gao, and Li 2022). Due to the class-agnostic feature, PCSOD can provide the pre-processing results for many downstream tasks, such as 3D shape classification, compression, and quality assessment.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

WeakPCSOD: Overcoming the Bias of Box Annotations for Weakly Supervised Point Cloud Salient Object Detection

Wei,

Zhou,

Cui

et al. 2024

AAAI

View full text Add to dashboard Cite

Point cloud salient object detection (PCSOD) is a newly proposed task in 3D dense segmentation. However, the acquisition of accurate 3D dense annotations comes at a high cost, severely limiting the progress of PCSOD. To address this issue, we propose the first weakly supervised PCSOD (named WeakPCSOD) model, which relies solely on cheap 3D bounding box annotations. In WeakPCSOD, we extract noise-free supervision from coarse 3D bounding boxes while mitigating shape biases inherent in box annotations. To achieve this, we introduce a novel mask-to-box (M2B) transformation and a color consistency (CC) loss. The M2B transformation, from a shape perspective, disentangles predictions from labels, enabling the extraction of noiseless supervision from labels while preserving object shapes independently of the box bias. From an appearance perspective, we further introduce the CC loss to provide dense supervision, which mitigates the non-unique predictions stemming from weak supervision and substantially reduces prediction variability. Furthermore, we employ a self-training (ST) strategy to enhance performance by utilizing high-confidence pseudo labels. Notably, the M2B transformation, CC loss, and ST strategy are seamlessly integrated into any model and incur no computational costs for inference. Extensive experiments demonstrate the effectiveness of our WeakPCSOD model, even comparable to fully supervised models utilizing dense annotations.

show abstract

“…2 shows the pipeline. For the specific structures of FAB, PPB, and SPB, please refer to (Fan, Gao, and Li 2022).…”

Section: Proposed Methodology Baseline Modelsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

WeakPCSOD: Overcoming the Bias of Box Annotations for Weakly Supervised Point Cloud Salient Object Detection

Wei,

Zhou,

Cui

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

Background Knowledge

Li,

Gao,

Gao

2024

Point Cloud Compression

View full text Add to dashboard Cite

Attention-guided LiDAR segmentation and odometry using image-to-point cloud saliency transfer

Ding,

İmamoğlu,

Caglayan

et al. 2024

Multimedia Systems

View full text Add to dashboard Cite

LiDAR odometry estimation and 3D semantic segmentation are crucial for autonomous driving, which has achieved remarkable advances recently. However, these tasks are challenging due to the imbalance of points in different semantic categories for 3D semantic segmentation and the influence of dynamic objects for LiDAR odometry estimation, which increases the importance of using representative/salient landmarks as reference points for robust feature learning. To address these challenges, we propose a saliency-guided approach that leverages attention information to improve the performance of LiDAR odometry estimation and semantic segmentation models. Unlike in the image domain, only a few studies have addressed point cloud saliency information due to the lack of annotated training data. To alleviate this, we first present a universal framework to transfer saliency distribution knowledge from color images to point clouds, and use this to construct a pseudo-saliency dataset (i.e. FordSaliency) for point clouds. Then, we adopt point cloud based backbones to learn saliency distribution from pseudo-saliency labels, which is followed by our proposed SalLiDAR module. SalLiDAR is a saliency-guided 3D semantic segmentation model that integrates saliency information to improve segmentation performance. Finally, we introduce SalLONet, a self-supervised saliency-guided LiDAR odometry network that uses the semantic and saliency predictions of SalLiDAR to achieve better odometry estimation. Our extensive experiments on benchmark datasets demonstrate that the proposed SalLiDAR and SalLONet models achieve state-of-the-art performance against existing methods, highlighting the effectiveness of image-to-LiDAR saliency knowledge transfer. Source code will be available at https://github.com/nevrez/SalLONet

show abstract

Salient Object Detection for Point Clouds

Cited by 12 publications

References 58 publications

WeakPCSOD: Overcoming the Bias of Box Annotations for Weakly Supervised Point Cloud Salient Object Detection

WeakPCSOD: Overcoming the Bias of Box Annotations for Weakly Supervised Point Cloud Salient Object Detection

Background Knowledge

Attention-guided LiDAR segmentation and odometry using image-to-point cloud saliency transfer

Contact Info

Product

Resources

About