PointMotionNet: Point-Wise Motion Learning for Large-Scale LiDAR Point Clouds Sequences

Wang, Jun; Li, Xiaolong; Sullivan, Alan; Abbott, Lynn; Chen, Siheng

doi:10.1109/cvprw56347.2022.00488

Cited by 9 publications

(11 citation statements)

References 35 publications

(64 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…46 Wang et al presented PointMotionNet, a spatiotemporal convolutional neural network that extracts spatio-temporal features, to distinguish the moving and static objects. 47 Pruim et al built upon the popular U-Net architecture by replacing 2D convolutions with 3D convolutions to detect maritime targets in video. The 3D U-Net spatio-temporal network outperforms a 2D U-Net that performs per-frame segmentation.…”

Section: Related Workmentioning

confidence: 99%

Leveraging temporal context in deep learning methodology for small object detection

Heslinga,

Ruis,

Ballan

et al. 2023

Artificial Intelligence for Security and Defence Applications

View full text Add to dashboard Cite

Automated object detection is becoming more relevant in a wide variety of applications in the military domain. This includes the detection of drones, ships, and vehicles in video and IR video. In recent years, deep learningbased object detection methods, such as YOLO, have shown to be promising in many applications for object detection. However, current methods have limited success when objects of interest are small in number of pixels, e.g. objects far away or small objects closer by. This is important, since accurate small object detection translates to early detection and the earlier an object is detected the more time is available for action. In this study, we investigate novel image analysis techniques that are designed to address some of the challenges of (very) small object detection by taking into account temporal information. We implement six methods, of which three are based on deep learning and use the temporal context of a set of frames within a video. The methods consider neighboring frames when detecting objects, either by stacking them as additional channels or by considering difference maps. We compare these spatio-temporal deep learning methods with YOLO-v8 that only considers single frames and two traditional moving object detection methods. Evaluation is done on a set of videos that encompasses a wide variety of challenges, including various objects, scenes, and acquisition conditions to show real-world performance.

show abstract

Section: Related Workmentioning

confidence: 99%

Leveraging temporal context in deep learning methodology for small object detection

Heslinga,

Ruis,

Ballan

et al. 2023

Artificial Intelligence for Security and Defence Applications

View full text Add to dashboard Cite

show abstract

“…Wang et al. proposed a point‐wise motion learning network (PointMotionNet) [27], which learns motion information from a sequence of large‐scale 3D LiDAR point clouds. Peri et al.…”

Section: Related Workmentioning

confidence: 99%

“…Salzmann et al [26] incorporated semantic maps and camera images for motion prediction. Wang et al proposed a point-wise motion learning network (PointMotionNet) [27], which learns motion information from a sequence of large-scale 3D LiDAR point clouds. Peri et al proposed a network that forecasts trajectories from LiDAR via future object detection [28].…”

Section: F I G U R E 1 a Comparison Of Long Short-term Memory (Lstm) Rawmentioning

confidence: 99%

Obstacle‐transformer: A trajectory prediction network based on surrounding trajectories

Zhang

Chai

Zhang

et al. 2022

IET Cyber-Syst and Robotics

View full text Add to dashboard Cite

Recurrent Neural Network, Long Short-Term Memory, and Transformer have made great progress in predicting the trajectories of moving objects. Although the trajectory element with the surrounding scene features has been merged to improve performance, there still exist some problems to be solved. One is that the time series processing models will increase the inference time with the increase of the number of prediction sequences. Another problem is that the features cannot be extracted from the scene's image and point cloud in some situations. Therefore, an Obstacle-Transformer is proposed to predict trajectory in a constant inference time. An 'obstacle' is designed by the surrounding trajectory rather than images or point clouds, making Obstacle-Transformer more applicable in a wider range of scenarios. Experiments are conducted on ETH and UCY datasets to verify the performance of our model.

show abstract

“…However, by processing the neighbourhoods at multiple scales, the network can capture temporal correlations that would otherwise be hidden. This hierarchical learning strategy has proved to be highly successful at learning from point cloud sequences and has been widely adopted throughout the literature [8,10,15,18,19,24,37,45]. In PSTNet [8] a hierarchical architecture is used for the action classification of point cloud sequences.…”

Section: Related Workmentioning

confidence: 99%

Do Users Behave Similarly in VR? Investigation of the User Influence on the System Design

Rossi

Özçinar

Smolić

et al. 2020

ACM Trans. Multimedia Comput. Commun. Appl.

View full text Add to dashboard Cite

With the overarching goal of developing user-centric Virtual Reality (VR) systems, a new wave of studies focused on understanding how users interact in VR environments has recently emerged. Despite the intense efforts, however, current literature still does not provide the right framework to fully interpret and predict users trajectories while navigating in VR scenes. This work advances the state-of-the-art on both the study of users' behaviour in VR and the user-centric system design. In more details, we complement current datasets by presenting a public available dataset that provides navigation trajectories acquired for heterogeneous omnidirectional videos and different viewing platforms, namely, head-mounted display, tablet and laptop. We then present an exhaustive analysis on the collected data, to better understand navigation in VR across users, content, and for the first time across viewing platforms. The novelty lies in the user-affinity metric, proposed in this work to investigate users' similarities when navigating within the content. The analysis reveals useful insights on the effect of device and content on the navigation, which could be precious considerations from the system design perspective. As a case study of the importance of studying users' behaviour when designing VR systems, we finally propose an user-centric server optimisation. We formulate an integer linear program that seeks the best stored set of omnidirectional content that minimises encoding and storage cost while maximises the user's experience. This is posed while taking into account network dynamics, type of video content, but also user population interactivity. Experimental results prove that our solution outperforms commonly company recommendations in terms of experienced quality but also in terms of encoding and storage, achieving a saving up to 70%. More importantly, we highlight a strong correlation between the storage cost and the user-affinity metric, showing the impact of the latter in the system architecture design.

show abstract

PointMotionNet: Point-Wise Motion Learning for Large-Scale LiDAR Point Clouds Sequences

Cited by 9 publications

References 35 publications

Leveraging temporal context in deep learning methodology for small object detection

Leveraging temporal context in deep learning methodology for small object detection

Obstacle‐transformer: A trajectory prediction network based on surrounding trajectories

Do Users Behave Similarly in VR? Investigation of the User Influence on the System Design

Contact Info

Product

Resources

About