Filters








15,476 Hits in 5.9 sec

Maximum Volume Subset Selection for Anchored Boxes

Karl Bringmann, Sergio Cabello, Michael Emmerich
unpublished
We study the problem of selecting k boxes in B that maximize the volume of the union of the selected boxes.  ...  The research is motivated by applications in skyline queries for databases and in multicriteria optimization, where the problem is known as the hypervolume subset selection problem.  ...  We are grateful to the other participants of the workshop and the Lorentz Center for their support. We are especially grateful to Günter Rote for several discussions and related work.  ... 
fatcat:y4pxqnlj65fq5o3cdl63gg5mba

Maximum Volume Subset Selection for Anchored Boxes [article]

Karl Bringmann, Sergio Cabello, Michael T.M. Emmerich
2018 pre-print
We study the problem of selecting k boxes in B that maximize the volume of the union of the selected boxes.  ...  This research is motivated by applications in skyline queries for databases and in multicriteria optimization, where the problem is known as the hypervolume subset selection problem.  ...  We are grateful to the other participants of the workshop and the Lorentz Center for their support. We are especially grateful to Günter Rote for several discussions and related work.  ... 
doi:10.4230/lipics.socg.2017.22 arXiv:1803.00849v1 fatcat:55prm7eonrcardtuhdlanlfj3u

Towards guided mutagenesis: Gaussian process regression predicts MHC class II antigen mutant binding [article]

David R. Bell, Serena H. Chen
2021 bioRxiv   pre-print
Using data from different residue subsets, we interpolate pMHCII mutant binding affinities by Gaussian process (GP) regression of residue volume and hydrophobicity.  ...  Our work finds that prediction is most accurate for neutral residues at anchor residue sites without register shift. This work holds relevance to predicting pMHCII binding and accelerating ASI design.  ...  Acknowledgements The authors would like to thank Leili Zhang, Guojing Cong, Giacomo Domeniconi, Chih-Chieh Yang, Ruhong Zhou, Jeffrey K Weber, and Sangyun Lee for insightful discussions.  ... 
doi:10.1101/2021.04.14.439878 fatcat:rtcn3swivbclvhubguh3ezqxha

Fragment oriented molecular shapes

Ethan Hain, Carlos J. Camacho, David Ryan Koes
2016 Journal of Molecular Graphics and Modelling  
FOMS enables the use of shape constraints, a novel method for  ...  We calculate a p-value for the returned subset (relative to the null hypothesis of random selection) using a hypergeometric test.  ...  For these queries the median time, as shown in the box plots, remains below a tenth of a second.  ... 
doi:10.1016/j.jmgm.2016.03.017 pmid:27085751 pmcid:PMC4862882 fatcat:3aj5zmakkrah3npvjqfqr47dua

SADet: Learning An Efficient and Accurate Pedestrian Detector [article]

Chubin Zhuang and Zhen Lei and Stan Z. Li
2020 arXiv   pre-print
Thirdly, we also design Cosine-NMS for the postprocess of predicted bounding boxes, and further propose adaptive anchor matching to enable the model to adaptively match the anchor boxes to full or visible  ...  bounding boxes according to the degree of occlusion, making the NMS and anchor matching algorithms more suitable for occluded pedestrian detection.  ...  Cosine-NMS Non-Maximum Suppression (NMS) is an integral part of the object detection pipeline, which recursively selects the detection box with the maximum score and remove the repeated predictions.  ... 
arXiv:2007.13119v1 fatcat:tpvarkvekjha3o3zpypj3gwtiy

Multi-view X-ray R-CNN [article]

Jan-Martin O. Steitz, Faraz Saeedan, Stefan Roth
2018 arXiv   pre-print
Motivated by the detection of prohibited objects in carry-on luggage as a part of avionic security screening, we develop a CNN-based object detection approach for multi-view X-ray image data.  ...  K-means clustering of anchor boxes When we expand the hand-selected aspect ratios of the Faster R-CNN anchor boxes of 1:1, 1:2, and 2:1 at 3 different scales to 3D, we arrive at a total of 21 anchor boxes  ...  Even a low number of clusters already outperforms the standard hand-selected anchor boxes (IoU of 0.5).  ... 
arXiv:1810.02344v1 fatcat:yu5vvh6orbdj5nijgdz2dc7rue

Hardness of discrepancy computation and ε-net verification in high dimension

Panos Giannopoulos, Christian Knauer, Magnus Wahlström, Daniel Werner
2012 Journal of Complexity  
As such a dependency on d becomes intractable for high-dimensional data, we ask whether it can be moderated.  ...  Depending on the ranges, several variants arise, including star discrepancy, box discrepancy, and discrepancy of halfspaces.  ...  Acknowledgments The authors thank the anonymous reviewers for their very helpful comments.  ... 
doi:10.1016/j.jco.2011.09.001 fatcat:oydluulgx5djbgv4m5qpur475i

Design, Analysis and Application of A Volumetric Convolutional Neural Network [article]

Xiaqing Pan, Yueru Chen, C.-C. Jay Kuo
2017 arXiv   pre-print
For the analysis of the VCNN, the cause of confusing classes in the output of the VCNN is explained by analyzing the relationship between the filter weights (also known as anchor vectors) from the last  ...  The proposed VCNN offers the state-of-the-art performance among all volume-based CNN methods.  ...  Acknowledgment Computation for the work described in this paper was supported by the University of Southern California's Center for High-Performance Computing (hpc.usc.edu).  ... 
arXiv:1702.00158v1 fatcat:dibzzyqwgvdfnh6kado3zxq7b4

Efficient Low-Cost Ship Detection for SAR Imagery Based on Simplified U-Net

Yuxing Mao, Yuqin Yang, Ziyuan Ma, Mingzhe Li, Hao Su, Jun Zhang
2020 IEEE Access  
INDEX TERMS Bounding box, score map, simplified U-Net, anchor-free, low-cost, SAR ship detection.  ...  Secondly, an anchor-free SAR ship detection framework consisting of a bounding boxes regression sub-net and a score map regression sub-net based on simplified U-Net is proposed.  ...  For more information, see https://creativecommons.org/licenses/by/4.0/ VOLUME 8, 2020 VOLUME 8, 2020  ... 
doi:10.1109/access.2020.2985637 fatcat:ydffxw65tnfszbudkmfo3c2q74

Focused Decoding Enables 3D Anatomical Detection by Transformers [article]

Bastian Wittmann, Fernando Navarro, Suprosanna Shit, Bjoern Menze
2022 arXiv   pre-print
Focused Decoder leverages information from an anatomical region atlas to simultaneously deploy query anchors and restrict the cross-attention's field of view to regions of interest, which allows for a  ...  Code for Focused Decoder is available in our medical Vision Transformer library github.com/bwittmann/transoar.  ...  Additionally, we estimate for each labeled anatomical structure the median, minimum, and maximum bounding box size, which will be necessary for the query anchor generation process and the concept of relative  ... 
arXiv:2207.10774v1 fatcat:gzcpdm72njdozllgs7cwhhpsrm

Occlusion Handling and Multi-Scale Pedestrian Detection Based on Deep Learning: A Review

Fang Li, Xueyuan Li, Qi Liu, Zirui Li
2022 IEEE Access  
Moreover, the popular datasets and evaluation methods for pedestrian detection are introduced. Finally, the development trend of pedestrian detection is prospected.  ...  selection.  ...  Results are obtained from original paper. 8 VOLUME 4, 2016 VOLUME 4, 2016 VOLUME 4, 2016 VOLUME 4, 2016 VOLUME 4, 2016 VOLUME 4, 2016  ... 
doi:10.1109/access.2022.3150988 fatcat:d6ym4kgy3jg5pa4eob7r3seuzu

On Guaranteed Optimal Robust Explanations for NLP Models [article]

Emanuele La Malfa, Agnieszka Zbrzezny, Rhiannon Michelmore, Nicola Paoletti, Marta Kwiatkowska
2021 arXiv   pre-print
We present two solution algorithms, respectively based on implicit hitting sets and maximum universal subsets, introducing a number of algorithmic improvements to speed up convergence of hard instances  ...  We build on abduction-based explanations for ma-chine learning and develop a method for computing local explanations for neural network models in natural language processing (NLP).  ...  which in turn a maximum universal subset (MUS) is required.  ... 
arXiv:2105.03640v2 fatcat:kc7tgbnhbre6hpnncllu4gad2u

Wide-Area Search Tracking for Siamese Region Proposal Network

Hongwei Zhang, Xiaoxia Li, Bin Zhu, Qi Ma
2020 IEEE Access  
We express our thanks for the experiment equipment provided by the lab.  ...  The former is a downsampled subset with a maximum sequence length of 1029 frames, and the latter is a subset containing 20 long-term sequences with an average sequence length of 2934 frames.  ...  UAV123_10fps and UAV20L are two subsets of the UAV123 benchmark.  ... 
doi:10.1109/access.2020.3003347 fatcat:kuzvzwxb2fgipgusn64hyerskm

Tube-CNN: Modeling temporal evolution of appearance for object detection in video [article]

Tuan-Hung Vu, Anton Osokin, Ivan Laptev
2018 arXiv   pre-print
To model temporal evolution, we introduce space-time tubes corresponding to temporal sequences of bounding boxes. We propose two CNN architectures for generating and classifying tubes, respectively.  ...  Our goal in this paper is to learn discriminative models for the temporal evolution of object appearance and to use such models for object detection.  ...  To study the detection performance under occlusions, we select a subset of video clips, named HollywoodHeads-Hard, with partially visible heads.  ... 
arXiv:1812.02619v1 fatcat:idogodsgsjf3xjxrwapzvvmk7i

Object as Hotspots: An Anchor-Free 3D Object Detection Approach via Firing of Hotspots [article]

Qi Chen, Lin Sun, Zhixin Wang, Kui Jia, Alan Yuille
2020 arXiv   pre-print
We thus argue in this paper for an approach opposite to existing methods using object-level anchors.  ...  Existing methods strive to organize the points regularly, e.g. voxelize, pass them through a designed 2D/3D neural network, and then define object-level anchors that predict offsets of 3D bounding boxes  ...  XYZ , Ernest Cheung (Samsung), Gweltaz Lever (Samsung), and Chenxu Luo (Johns Hopkins University and Samsung) for useful discussions that greatly improved the manuscript.  ... 
arXiv:1912.12791v3 fatcat:ze7rt6zr5jajjbbhd42kpnmdva
« Previous Showing results 1 — 15 out of 15,476 results