Filters








18,561 Hits in 2.7 sec

Neural Scene Graphs for Dynamic Scenes [article]

Julian Ost, Fahim Mannan, Nils Thuerey, Julian Knodt, Felix Heide
2021 arXiv   pre-print
In this work, we present the first neural rendering method that decomposes dynamic scenes into scene graphs.  ...  However, existing methods are restricted to learning efficient representations of static scenes that encode all scene objects into a single neural network, and lack the ability to represent dynamic scenes  ...  Our neural scene graph method achieves high-quality view synthesis results regardless of this dynamicity, both for dynamic scene reconstruction and novel scene generation.  ... 
arXiv:2011.10379v3 fatcat:k4lytj7clbewbjkqoao7e2js5e

Visual Experience-Based Question Answering with Complex Multimodal Environments

Incheol Kim, Jiayi Ma
2020 Mathematical Problems in Engineering  
The proposed system can generate more accurate scene graphs for dynamic environments with some uncertainty.  ...  To address this VEQA problem, we propose a hybrid visual question answering system, VQAS, integrating a deep neural network-based scene graph generation model and a rule-based knowledge reasoning system  ...  scene graph for answer generation.  ... 
doi:10.1155/2020/8567271 fatcat:cgmzylh4ujadfisikt5hekvrga

Neural Scene Flow Prior [article]

Xueqian Li, Jhony Kaesemodel Pontes, Simon Lucey
2021 arXiv   pre-print
A central innovation here is the inclusion of a neural scene flow prior, which uses the architecture of neural networks as a new type of implicit regularizer.  ...  A prime example of this in computer vision is optical and scene flow. Supervised learning has largely displaced the need for explicit regularization.  ...  We thank Haosen Xing for careful review of the entire manuscript and assistance in several parts of the paper, Jianqiao Zheng for helpful discussions.  ... 
arXiv:2111.01253v1 fatcat:tl3afmyqjraxxilsarfghiehqe

A NEURAL SYSTEM FOR THE RECOGNITION OF PARTIALLY OCCLUDED OBJECTS IN CLUTTERED SCENES: A PILOT STUDY

LAURENZ WISKOTT, CHRISTOPH VON DER MALSBURG
1993 International journal of pattern recognition and artificial intelligence  
The system architecture goes beyond classical neural networks by making extensive use of flexible links between units, as proposed in the dynamic link architecture.  ...  We present a system for the interpretation of camera images of scenes composed of several known objects with mutual occlusion.  ...  Acknowledgements: We wish to thank Martin Lades for his contributions to the development of the program package and Wolfgang Konen for useful discussions.  ... 
doi:10.1142/s0218001493000479 fatcat:c3b3dcm3freqdmiuw4owb353ky

A NEURAL SYSTEM FOR THE RECOGNITION OF PARTIALLY OCCLUDED OBJECTS IN CLUTTERED SCENES: A PILOT STUDY [chapter]

LAURENZ WISKOTT, CHRISTOPH VON DER MALSBURG
1994 Series in Machine Perception and Artificial Intelligence  
The system architecture goes beyond classical neural networks by making extensive use of flexible links between units, as proposed in the dynamic link architecture.  ...  We present a system for the interpretation of camera images of scenes composed of several known objects with mutual occlusion.  ...  Acknowledgements: We wish to thank Martin Lades for his contributions to the development of the program package and Wolfgang Konen for useful discussions.  ... 
doi:10.1142/9789812797926_0017 fatcat:ddrettfvoff6pbg6pquqk5jhcq

TöRF: Time-of-Flight Radiance Fields for Dynamic Scene View Synthesis [article]

Benjamin Attal, Eliot Laidlaw, Aaron Gokaslan, Changil Kim, Christian Richardt, James Tompkin, Matthew O'Toole
2021 arXiv   pre-print
Neural networks can represent and accurately reconstruct radiance fields for static 3D scenes (e.g., NeRF).  ...  Several works extend these to dynamic scenes captured with monocular video, with promising performance.  ...  Further, neural networks have limited capacity to model dynamic scenes, which limits the duration of dynamic sequences. This is a limitation of many current neural dynamic scene methods. 6.2.  ... 
arXiv:2109.15271v2 fatcat:bxm73ltkobaivjrxc4yv2izsoy

Continuous Graph Flow [article]

Zhiwei Deng, Megha Nawhal, Lili Meng, Greg Mori
2019 arXiv   pre-print
This leads to a new type of neural graph message passing scheme that performs continuous message passing over time.  ...  We demonstrate the effectiveness of our model on a diverse set of generation tasks across different domains: graph generation, image puzzle generation, and layout generation from scene graphs.  ...  Layout generation for scene graphs.  ... 
arXiv:1908.02436v2 fatcat:wq2xgqbvbvcdzgs5k2abubge5e

Residual Attention Graph Convolutional Network for Geometric 3D Scene Classification [article]

Albert Mosella-Montoro, Javier Ruiz-Hidalgo
2019 arXiv   pre-print
Experimental results show that the proposed method outperforms current state-of-the-art in geometric 3D scene classification tasks.  ...  This work proposes a novel Residual Attention Graph Convolutional Network that exploits the intrinsic geometric context inside a 3D space without using any kind of point features, allowing the use of organized  ...  The main contributions of this paper are: a) The use of an Attention Graph Neural Network to capture the geometric information of a scene. b) The adaptation of Attention Graph Neural Networks into a deep  ... 
arXiv:1909.13470v1 fatcat:eemi3u65vbamlnzeuyx4fpwvwi

Hierarchical Planning for Long-Horizon Manipulation with Geometric and Symbolic Scene Graphs [article]

Yifeng Zhu, Jonathan Tremblay, Stan Birchfield, Yuke Zhu
2021 arXiv   pre-print
Our model uses graph neural networks to process these scene graphs for predicting high-level task plans and low-level motions.  ...  At the core of our approach is a two-level scene graph representation, namely geometric scene graph and symbolic scene graph.  ...  ACKNOWLEDGEMENT We would like to thank Guanya Shi for providing an improved version of DOPE. We also would like to thank Nvidia AI-Algorithm team for providing valuable internal feedback to the paper.  ... 
arXiv:2012.07277v2 fatcat:fejesoya35cubauvnjsbrjnqjq

Grouptron: Dynamic Multi-Scale Graph Convolutional Networks for Group-Aware Dense Crowd Trajectory Forecasting [article]

Rui Zhou, Hongyu Zhou, Masayoshi Tomizuka, Jiachen Li, Zhuo Xu
2021 arXiv   pre-print
It then uses graph neural networks to encode dynamics at different scales and incorporates encoding across different scales for trajectory prediction.  ...  In this paper, we present Grouptron, a multi-scale dynamic forecasting framework that leverages pedestrian group detection and utilizes individual-level, group-level, and scene-level information for better  ...  Concretely, we construct a dynamic multi-scale graph convolutional neural network that uses pedestrian group information to better learn and model the extremely complex dynamics at different scales and  ... 
arXiv:2109.14128v1 fatcat:v7daz4z2svhe3o5xchjjt46arm

LRTA: A Transparent Neural-Symbolic Reasoning Framework with Modular Supervision for Visual Question Answering [article]

Weixin Liang, Feiyang Niu, Aishwarya Reganti, Govind Thattai, Gokhan Tur
2020 arXiv   pre-print
It then executes the reasoning instructions one at a time by traversing the scene graph using a recurrent neural-symbolic execution module.  ...  Specifically, LRTA learns to first convert an image into a scene graph and parse a question into multiple reasoning instructions.  ...  Acknowledgement We would like to thank Robinson Piramuthu, Dilek Hakkani-Tur, Arindam Mandal, Yanbang Wang and the anonymous reviewers for their insightful feedback and discussions that have notably shaped  ... 
arXiv:2011.10731v1 fatcat:jk7gpqjhvjdpnosoyc6ccpjfde

Page 1205 of Neural Computation Vol. 8, Issue 6 [page]

1996 Neural Computation  
A graph representation of a visual scene can consist of graph nodes that code local scene attributes, and graph arcs that code relationships and control communication between them, useful for relaxation  ...  For a neural network to create such a graph representation, certain implementational constraints need to be realized.  ... 

Attentive Gated Graph Neural Network for Image Scene Graph Generation

Shuohao Li, Min Tang, Jun Zhang, Lincheng Jiang
2020 Symmetry  
In this work, we translate the scene graph into an Attentive Gated Graph Neural Network which can propagate a message by visual relationship embedding.  ...  Image scene graph is a semantic structural representation which can not only show what objects are in the image, but also infer the relationships and interactions among them.  ...  Acknowledgments: The authors acknowledge Professor Takayuki Okatani and Professor Kota Yamaguchi in Tohuku University for their valuable suggestions which improved the results and presentation of this  ... 
doi:10.3390/sym12040511 fatcat:5v57eu723bglpm44pib6xj2fje

Hierarchical Representations and Explicit Memory: Learning Effective Navigation Policies on 3D Scene Graphs using Graph Neural Networks [article]

Zachary Ravichandran, Lisa Peng, Nathan Hughes, J. Daniel Griffith, Luca Carlone
2021 arXiv   pre-print
Towards this goal, we propose a graph neural network architecture and show how to embed a 3D scene graph into an agent-centric feature space, which enables the robot to learn policies for low-level action  ...  For each node in the scene graph, our method uses features that capture occupancy and semantic content, while explicitly retaining memory of the robot trajectory.  ...  Graph Neural Networks for Reinforcement Learning.  ... 
arXiv:2108.01176v1 fatcat:x6l4k6tzyjblbo2f222lby7vvi

Scene Context Based Semantic Segmentation for 3D LiDAR Data in Dynamic Scene [article]

Jilin Mei, Huijing Zhao
2020 arXiv   pre-print
We propose a graph neural network(GNN) based method to incorporate scene context for the semantic segmentation of 3D LiDAR data.  ...  Quantitative evaluation on a dataset of dynamic scene shows that our method has better performance than unary CNN with 8% improvement, as well as normal GNN with 17% improvement.  ...  The graph neural network(GNN) [7] provides a basic solution for our task.  ... 
arXiv:2003.13926v1 fatcat:shlrmskv55dhjjckyiwptc5cge
« Previous Showing results 1 — 15 out of 18,561 results