A Hierarchical Graph Network for 3D Object Detection on Point Clouds

Chen, Jintai; Lei, Biwen; Song, Qingyu; Ying, Haochao; Chen, Danny Z.; Wu, Jian

doi:10.1109/cvpr42600.2020.00047

Cited by 130 publications

(76 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, many PC-based object detectors do not effectively capture the semantic characteristic of PCs. In this direction, Chen et al [71] introduced the hierarchical graph network (HGNet) as shown in Figure 18 that processes raw PCs using multi-level semantics for 3D object detection. It contained three main parts, which are a graph convolution-based U-shape network called GUnet, proposal generator, and proposal reasoning module (referred to as ProRe Module).…”

Section: (Iii) Graph Representation For 3dormentioning

confidence: 99%

“…HGNet [71] o o S-AT GCN [72] o Table 18 MV3D [73] o BEVLFVC [74] o D3PD [75] o MVX-Net [76] o SharedNet [77] o 3DPR Several methods discussed in the survey illustrate that KITTI dataset [172] published in 2012 by [173] is the most frequently used dataset for 3DOR tasks. The review shows that many 3DOR models (19 out of 23 studies) have used the KITTI dataset.…”

Section: Datasetsmentioning

confidence: 99%

“…The SUN RGB-D dataset [182] used by one 3DPR [78] and two 3DOR [70,71] studies was presented by [183]. It contains 10,355 RGB-D scene images as training set and 2860 images as testing set for 3D object detection, which is fundamental for scene understating.…”

Section: Datasetsmentioning

confidence: 99%

See 2 more Smart Citations

3D Recognition Based on Sensor Modalities for Robotic Systems: A Survey

Manzoor

Joo

Kim

et al. 2021

Sensors

View full text Add to dashboard Cite

3D visual recognition is a prerequisite for most autonomous robotic systems operating in the real world. It empowers robots to perform a variety of tasks, such as tracking, understanding the environment, and human–robot interaction. Autonomous robots equipped with 3D recognition capability can better perform their social roles through supportive task assistance in professional jobs and effective domestic services. For active assistance, social robots must recognize their surroundings, including objects and places to perform the task more efficiently. This article first highlights the value-centric role of social robots in society by presenting recently developed robots and describes their main features. Instigated by the recognition capability of social robots, we present the analysis of data representation methods based on sensor modalities for 3D object and place recognition using deep learning models. In this direction, we delineate the research gaps that need to be addressed, summarize 3D recognition datasets, and present performance comparisons. Finally, a discussion of future research directions concludes the article. This survey is intended to show how recent developments in 3D visual recognition based on sensor modalities using deep-learning-based approaches can lay the groundwork to inspire further research and serves as a guide to those who are interested in vision-based robotics applications.

show abstract

Section: (Iii) Graph Representation For 3dormentioning

confidence: 99%

Section: Datasetsmentioning

confidence: 99%

See 1 more Smart Citation

3D Recognition Based on Sensor Modalities for Robotic Systems: A Survey

Manzoor

Joo

Kim

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…On other hand, in [31] used depth and colour data to determine automatically the location of an object and minimized the difficulty of visual analysis by using salient item recognition for RGB-D pictures. They proposed salient item recognition by convolution neural network with a single stream.…”

Section: Literature Reviewmentioning

confidence: 99%

Hybrid features for object detection in RGB-D scenes

Awwad

Igried

Wedyan

et al. 2021

IJEECS

View full text Add to dashboard Cite

<div>Object detection is considered a hot research topic in applications of artificial intel-ligence and computer vision. Historically, object detection was widely used in var-ious fields like surveillance, fine-grained activities and robotics. All studies focus on improving accuracy for object detection using images, whether indoor or outdoor scenes. Therefore, this paper took a shot by improving the doable features extraction and proposing crossed sliding window approach using exiting classifiers for object de-tection. In this paper, the contribution includes two parts: First, improving local depth pattern feature along side SIFT and the second part explains a new technique presented by proposing crossed sliding window approach using two different types of images (colored and depth). Two types of features local depth patterns for detection (LDPD) and scale-invariant feature transform (SIFT) were merged as one feature vector. The RGB-D object dataset has been used and it consists of 300 different objects and in-cludes thousands of scenes. The proposed approach achieved high results comparing to other features or separated features that are used in this paper. All experiments and comparatives were applied on the same dataset for the same objective. Experimental results report a high accuracy in terms of detection rate, recall, precision and F1 scorein RGB-D scenes.</div>

show abstract

“…The third app shares the AR point cloud, which contains the 3D visual corner points that are used to track the space. AR apps can share point clouds among users for shared positioning and/or send it to the cloud for object detection [13] and/or image-based localization [58]. The fourth app shares the face tracking result.…”

Section: Benchmark Applicationsmentioning

confidence: 99%

LensCap

Iosifescu

LiKamWa

2021

Proceedings of the 19th Annual International Conference on Mobile Systems, Applications, and Services

View full text Add to dashboard Cite

Augmented Reality (AR) enables smartphone users to interact with virtual content spatially overlaid on a continuously captured physical world. Under the current permission enforcement model in popular operating systems, AR apps are given Internet permission at installation time, and request camera permission and external storage write permission at runtime through a user's approval. With these permissions granted, any Internet-enabled AR app could silently collect camera frames and derived visual information for malicious intent without a user's awareness. This raises serious concerns about the disclosure of private user data in their living environments.To give users more control over application usage of their camera frames and the information derived from them, we introduce LensCap, a split-process app design framework, in which the app is split into a camera-handling visual process and a connectivityhandling network process. At runtime, LensCap manages secured communications between split processes, enacting fine-grained data usage monitoring. LensCap also allows both processes to present interactive user interfaces. With LensCap, users can decide what forms of visual data can be transmitted to the network, while still allowing visual data to be used for AR purposes on device. We prototype LensCap as an Android library and demonstrate its usability as a plugin in Unreal Engine. Performance evaluation results on five AR apps confirm that visual privacy can be preserved with an insignificant latency penalty (< 1.3 ms) at 60 FPS. CCS CONCEPTS• Security and privacy → Software security engineering.

show abstract

A Hierarchical Graph Network for 3D Object Detection on Point Clouds

Cited by 130 publications

References 30 publications

3D Recognition Based on Sensor Modalities for Robotic Systems: A Survey

3D Recognition Based on Sensor Modalities for Robotic Systems: A Survey

Hybrid features for object detection in RGB-D scenes

LensCap

Contact Info

Product

Resources

About