A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Filters
Learning Direct Optimization for Scene Understanding
[article]
2018
arXiv
pre-print
Preserved Fulltext
Other Versions
We introduce a Learning Direct Optimization method for the refinement of a latent variable model that describes input image x. ...
Our novel Learning Direct Optimization (LiDO) approach trains a Prediction Network to predict an update directly to correct z, rather than minimizing the error with respect to z. ...
Supplementary Material for: Learning Direct Optimization for Scene Understanding Detector and Initialization Networks . . . . . 13 D.2. LiDO Prediction Network . . . . . . . . . . 13 A. ...
arXiv:1812.07524v1
fatcat:bgnahxc24jdwpavvjdmitiiiei
DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
image quality for total panoramic 3D scene understanding. ...
In this paper, we propose a novel method for panoramic 3D scene understanding which recovers the 3D room layout and the shape, pose, position, and semantic category for each object from a single full-view ...
To learn and evaluate our model, a new dataset is created for total panoramic 3D scene understanding. ...
arXiv:2108.10743v1
fatcat:knc6p65etvfihcb6rey3oihrfu
Towards Holistic Scene Understanding: Feedback Enabled Cascaded Classification Models
[article]
2011
arXiv
pre-print
Preserved Fulltext
The Internet Archive has a preservation copy of this work in our general collections.
The file type is application/pdf.
Scene understanding includes many related sub-tasks, such as scene categorization, depth estimation, object detection, etc. ...
We propose Feedback Enabled Cascaded Classification Models (FE-CCM), that jointly optimizes all the sub-tasks, while requiring only a 'black-box' interface to the original classifier for each sub-task. ...
ACKNOWLEDGMENTS We thank Anish Nahar, Matthew Cong, TP Wong, Norris Xu, and Colin Ponce for help with the robotic experiments. We also thank Daphne Koller for useful discussions. ...
arXiv:1110.5102v1
fatcat:abtffcuzkvftbhk5s3yi2og7y4
Holistic++ Scene Understanding: Single-view 3D Holistic Scene Parsing and Human Pose Estimation with Human-Object Interaction and Physical Commonsense
[article]
2019
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We propose a new 3D holistic++ scene understanding problem, which jointly tackles two tasks from a single-view image: (i) holistic scene parsing and reconstruction---3D estimations of object bounding boxes ...
The intuition behind is to leverage the coupled nature of these two tasks to improve the granularity and performance of scene understanding. ...
. • E hoi models HOI and provides strong and fine-grained constraints for holistic scene understanding. ...
arXiv:1909.01507v1
fatcat:svbd33j7hvaz5jbysjwwhqhnoy
Convolution Neural Network Architecture Learning for Remote Sensing Scene Classification
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
In this paper, we proposed an automatically architecture learning procedure for remote sensing scene classification. ...
To discover the optimal set of parameters for a given dataset, we introduced a learning strategy which can allow efficient search in the architecture space by means of gradient descent. ...
Architecture determines the function of CNN, this architecture-learned method may help to understand which representations are important for remote sensing scene classification tasks.
II. ...
arXiv:2001.09614v1
fatcat:drfzcrelbrhgngujdgrk3uu2ee
Holistic++ Scene Understanding: Single-View 3D Holistic Scene Parsing and Human Pose Estimation With Human-Object Interaction and Physical Commonsense
2019
2019 IEEE/CVF International Conference on Computer Vision (ICCV)
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We propose a new 3D holistic ++ scene understanding problem, which jointly tackles two tasks from a single-view image: (i) holistic scene parsing and reconstruction-3D estimations of object bounding boxes ...
The intuition behind is to leverage the coupled nature of these two tasks to improve the granularity and performance of scene understanding. ...
All the evidence calls for a treatment to integrate HOI and physical commonsense with a modern computer vision system for scene understanding. ...
doi:10.1109/iccv.2019.00874
dblp:conf/iccv/ChenHYZQZ19
fatcat:2dutdksvzjgind7okkv7puymmm
Neural Ray-Tracing: Learning Surfaces and Reflectance for Relighting and View Synthesis
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2021 pre-print arXiv:2104.13562v1 fatcat:wjwawueqhrbyffkqd4cpvyfjj4
We validate the proposed approach for scene editing, relighting and reflectance estimation learned from synthetic and captured views on a subset of NeRV's datasets. ...
The method outperforms NeRV for relighting under known lighting conditions, and produces realistic reconstructions for relit and edited scenes. ...
We also demonstrate the understandibility of our model, by showing surface normals and incoming light directions along the surface, by modifying integrators. ...
arXiv:2104.13562v2
fatcat:bgy25vm77jaudfcx4l4ftvwvfq
Insights into few shot learning approaches for image scene classification
2021
PeerJ Computer Science
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Notwithstanding the optimal performance of exiting machine learning models in image understanding and scene classification, there are still obstacles to overcome. ...
Motivated by these findings, in this paper, we introduce two models for few-shot learning in scene classification. ...
Despite the optimal performance of existing models in image understanding and scene classification, there are still major issues. ...
doi:10.7717/peerj-cs.666
pmid:34616882
pmcid:PMC8459776
fatcat:uzwb5qlf75ba3as6jwu4oaoukq
An Overview of Perception and Decision-Making in Autonomous Systems in the Era of Learning
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Finally, we examine the several challenges and promising directions discussed and concluded in related research for future works in the era of computer science, automatic control, and robotics. ...
First, we delineate the existing classical simultaneous localization and mapping (SLAM) solutions and review the environmental perception and understanding methods based on deep learning, including deep ...
It is a development direction for autonomous systems to realize the optimal strategy for task assignment ac-cording to the current state of robots and task requirements. ...
arXiv:2001.02319v3
fatcat:z3zhp2cyonfqtlttl2y57572uy
Deep Learning Driven Visual Path Prediction From a Single Image
2016
IEEE Transactions on Image Processing
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2016 pre-print arXiv:1601.07265v1 fatcat:lxpdkpwm4zc2vmzxi5hizizxtm
Its goal is to infer the future path for a visual object in a static scene. ...
Motivated by these observations, we propose a deep learning framework which simultaneously performs deep feature learning for visual representation in conjunction with spatio-temporal context modeling. ...
Which one is optimal? P 3 wins out because its initial direction θ P3 is closer to θ esti . ...
doi:10.1109/tip.2016.2613686
pmid:28114063
fatcat:vnvwbkv2vvhodbvha3jjuuzwqu
Panoptic Neural Fields: A Semantic Object-Aware Neural Scene Representation
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
During experiments with real-world dynamic scenes, we find that our model can be used effectively for several tasks like novel view synthesis, 2D panoptic segmentation, 3D scene editing, and multiview ...
Each object is represented by an oriented 3D bounding box and a multi-layer perceptron (MLP) that takes position, direction, and time and outputs density and radiance. ...
We use Adam as our optimizer for all our training routines including optimizing Panoptic Neural Field model to a new scene and also during inner loops of FedAvg based meta-learning (see Sec. 3.4) for learned ...
arXiv:2205.04334v1
fatcat:ukrp6g5wurbblmj6pbjlclu2im
Convex Optimization for Scene Understanding
2013
2013 IEEE International Conference on Computer Vision Workshops
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
In this paper we give a convex optimization approach for scene understanding. ...
Since segmentation, object recognition and scene labeling strongly benefit from each other we propose to solve these tasks within a single convex optimization problem. ...
Solving these tasks within a single optimization problem such that all tasks can influence each other improves results for scene understanding. ...
doi:10.1109/iccvw.2013.131
dblp:conf/iccvw/SouiaiNSC13
fatcat:4svrxguclbca5ezvsfn6z2upmy
GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose
[article]
2018
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Other Versions
We propose GeoNet, a jointly unsupervised learning framework for monocular depth, optical flow and ego-motion estimation from videos. ...
The three components are coupled by the nature of 3D scene geometry, jointly learned by our framework in an end-to-end manner. ...
Acknowledgements We would like to thank Guorun Yang and Tinghui Zhou for helpful discussions and sharing the code. We also thank the anonymous reviewers for their instructive comments. ...
arXiv:1803.02276v2
fatcat:rcqfxt53qzfehb2uiqoav7w42e
GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose
2018
2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
We propose GeoNet, a jointly unsupervised learning framework for monocular depth, optical flow and egomotion estimation from videos. ...
The three components are coupled by the nature of 3D scene geometry, jointly learned by our framework in an end-to-end manner. ...
Supervised Deep Models for Geometry Understanding With recent development of deep learning, great progress has been made in many tasks of 3D geometry understanding, including depth, optical flow, pose ...
doi:10.1109/cvpr.2018.00212
dblp:conf/cvpr/YinS18
fatcat:osotve7hgvetdgp6d5hvv2u32q
Developing an Interactive 3D Learning Experience to Help Students Understand Key Regulatory Processes Associated with Glycolysis
2018
Journal of Biocommunication
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Recent recommendations for undergraduate biologyinstruction emphasize teaching foundational biologicalprinciples and helping students transfer these principles tomore complex biological phenomena. ...
The following researchproject was developed to explore whether a visualinteractive didactic tool could be developed to improvelearning outcomes for undergraduate biology students onthe topic of allosteric ...
The finished application and 2D animation were optimized and exported for web delivery (see figures 6 and 7). The landing page contains a brief introduction along with the 2D animation. ...
doi:10.5210/jbc.v42i1.9193
fatcat:pyddxqm5xfamdazxw3pqhlnwqe
« Previous
Showing results 1 — 15 out of 105,271 results