4 Search Results for "Zeng, Ji"

Document
DOI: 10.4230/LIPIcs.SoCG.2023.59
On Higher Dimensional Point Sets in General Position

Authors: Andrew Suk and Ji Zeng

Published in: LIPIcs, Volume 258, 39th International Symposium on Computational Geometry (SoCG 2023)

Abstract
A finite point set in ℝ^d is in general position if no d + 1 points lie on a common hyperplane. Let α_d(N) be the largest integer such that any set of N points in ℝ^d with no d + 2 members on a common hyperplane, contains a subset of size α_d(N) in general position. Using the method of hypergraph containers, Balogh and Solymosi showed that α₂(N) < N^{5/6 + o(1)}. In this paper, we also use the container method to obtain new upper bounds for α_d(N) when d ≥ 3. More precisely, we show that if d is odd, then α_d(N) < N^{1/2 + 1/(2d) + o(1)}, and if d is even, we have α_d(N) < N^{1/2 + 1/(d-1) + o(1)}. We also study the classical problem of determining the maximum number a(d,k,n) of points selected from the grid [n]^d such that no k + 2 members lie on a k-flat. For fixed d and k, we show that a(d,k,n)≤ O(n^{d/{2⌊(k+2)/4⌋}(1- 1/{2⌊(k+2)/4⌋d+1})}), which improves the previously best known bound of O(n^{d/⌊(k + 2)/2⌋}) due to Lefmann when k+2 is congruent to 0 or 1 mod 4.
Cite as

Andrew Suk and Ji Zeng. On Higher Dimensional Point Sets in General Position. In 39th International Symposium on Computational Geometry (SoCG 2023). Leibniz International Proceedings in Informatics (LIPIcs), Volume 258, pp. 59:1-59:13, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2023)

Copy BibTex To Clipboard

@InProceedings{suk_et_al:LIPIcs.SoCG.2023.59,
  author =	{Suk, Andrew and Zeng, Ji},
  title =	{{On Higher Dimensional Point Sets in General Position}},
  booktitle =	{39th International Symposium on Computational Geometry (SoCG 2023)},
  pages =	{59:1--59:13},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-273-0},
  ISSN =	{1868-8969},
  year =	{2023},
  volume =	{258},
  editor =	{Chambers, Erin W. and Gudmundsson, Joachim},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.SoCG.2023.59},
  URN =		{urn:nbn:de:0030-drops-179097},
  doi =		{10.4230/LIPIcs.SoCG.2023.59},
  annote =	{Keywords: independent sets, hypergraph container method, generalised Sidon sets}
}
Document
DOI: 10.4230/LITES.8.1.1
Susceptibility to Image Resolution in Face Recognition and Training Strategies to Enhance Robustness

Authors: Martin Knoche, Stefan Hörmann, and Gerhard Rigoll

Published in: LITES, Volume 8, Issue 1 (2022): Special Issue on Embedded Systems for Computer Vision. Leibniz Transactions on Embedded Systems, Volume 8, Issue 1

Abstract
Many face recognition approaches expect the input images to have similar image resolution. However, in real-world applications, the image resolution varies due to different image capture mechanisms or sources, affecting the performance of face recognition systems. This work first analyzes the image resolution susceptibility of modern face recognition. Face verification on the very popular LFW dataset drops from 99.23% accuracy to almost 55% when image dimensions of both images are reduced to arguable very poor resolution. With cross-resolution image pairs (one HR and one LR image), face verification accuracy is even worse. This characteristic is investigated more in-depth by analyzing the feature distances utilized for face verification. To increase the robustness, we propose two training strategies applied to a state-of-the-art face recognition model: 1) Training with 50% low resolution images within each batch and 2) using the cosine distance loss between high and low resolution features in a siamese network structure. Both methods significantly boost face verification accuracy for matching training and testing image resolutions. Training a network with different resolutions simultaneously instead of adding only one specific low resolution showed improvements across all resolutions and made a single model applicable to unknown resolutions. However, models trained for one particular low resolution perform better when using the exact resolution for testing. We improve the face verification accuracy from 96.86% to 97.72% on the popular LFW database with uniformly distributed image dimensions between 112 × 112 px and 5 × 5 px. Our approaches improve face verification accuracy even more from 77.56% to 87.17% for distributions focusing on lower images resolutions. Lastly, we propose specific image dimension sets focusing on high, mid, and low resolution for five well-known datasets to benchmark face verification accuracy in cross-resolution scenarios.
Cite as

Martin Knoche, Stefan Hörmann, and Gerhard Rigoll. Susceptibility to Image Resolution in Face Recognition and Training Strategies to Enhance Robustness. In LITES, Volume 8, Issue 1 (2022): Special Issue on Embedded Systems for Computer Vision. Leibniz Transactions on Embedded Systems, Volume 8, Issue 1, pp. 01:1-01:20, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2022)

Copy BibTex To Clipboard

@Article{knoche_et_al:LITES.8.1.1,
  author =	{Knoche, Martin and H\"{o}rmann, Stefan and Rigoll, Gerhard},
  title =	{{Susceptibility to Image Resolution in Face Recognition and Training Strategies to Enhance Robustness}},
  booktitle =	{LITES, Volume 8, Issue 1 (2022): Special Issue on Embedded Systems for Computer Vision},
  pages =	{01:1--01:20},
  journal =	{Leibniz Transactions on Embedded Systems},
  ISSN =	{2199-2002},
  year =	{2022},
  volume =	{8},
  number =	{1},
  editor =	{Knoche, Martin and H\"{o}rmann, Stefan and Rigoll, Gerhard},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LITES.8.1.1},
  doi =		{10.4230/LITES.8.1.1},
  annote =	{Keywords: recognition, resolution, cross, face, identification}
}
Document
DOI: 10.4230/LIPIcs.SoCG.2022.62
A Positive Fraction Erdős-Szekeres Theorem and Its Applications

Authors: Andrew Suk and Ji Zeng

Published in: LIPIcs, Volume 224, 38th International Symposium on Computational Geometry (SoCG 2022)

Abstract
A famous theorem of Erdős and Szekeres states that any sequence of n distinct real numbers contains a monotone subsequence of length at least √n. Here, we prove a positive fraction version of this theorem. For n > (k-1)², any sequence A of n distinct real numbers contains a collection of subsets A_1,…, A_k ⊂ A, appearing sequentially, all of size s = Ω(n/k²), such that every subsequence (a_1,…, a_k), with a_i ∈ A_i, is increasing, or every such subsequence is decreasing. The subsequence S = (A_1,…, A_k) described above is called block-monotone of depth k and block-size s. Our theorem is asymptotically best possible and follows from a more general Ramsey-type result for monotone paths, which we find of independent interest. We also show that for any positive integer k, any finite sequence of distinct real numbers can be partitioned into O(k²log k) block-monotone subsequences of depth at least k, upon deleting at most (k-1)² entries. We apply our results to mutually avoiding planar point sets and biarc diagrams in graph drawing.
Cite as

Andrew Suk and Ji Zeng. A Positive Fraction Erdős-Szekeres Theorem and Its Applications. In 38th International Symposium on Computational Geometry (SoCG 2022). Leibniz International Proceedings in Informatics (LIPIcs), Volume 224, pp. 62:1-62:15, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2022)

Copy BibTex To Clipboard

@InProceedings{suk_et_al:LIPIcs.SoCG.2022.62,
  author =	{Suk, Andrew and Zeng, Ji},
  title =	{{A Positive Fraction Erd\H{o}s-Szekeres Theorem and Its Applications}},
  booktitle =	{38th International Symposium on Computational Geometry (SoCG 2022)},
  pages =	{62:1--62:15},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-227-3},
  ISSN =	{1868-8969},
  year =	{2022},
  volume =	{224},
  editor =	{Goaoc, Xavier and Kerber, Michael},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.SoCG.2022.62},
  URN =		{urn:nbn:de:0030-drops-160703},
  doi =		{10.4230/LIPIcs.SoCG.2022.62},
  annote =	{Keywords: Erd\H{o}s-Szekeres, block-monotone, monotone biarc diagrams, mutually avoiding sets}
}
Document
DOI: 10.4230/LIPIcs.TQC.2013.192
Symmetries of Codeword Stabilized Quantum Codes

Authors: Salman Beigi, Jianxin Chen, Markus Grassl, Zhengfeng Ji, Qiang Wang, and Bei Zeng

Published in: LIPIcs, Volume 22, 8th Conference on the Theory of Quantum Computation, Communication and Cryptography (TQC 2013)

Abstract
Symmetry is at the heart of coding theory. Codes with symmetry, especially cyclic codes, play an essential role in both theory and practical applications of classical error-correcting codes. Here we examine symmetry properties for codeword stabilized (CWS) quantum codes, which is the most general framework for constructing quantum error-correcting codes known to date. A CWS code Q can be represented by a self-dual additive code S and a classical code C, i.e., Q=(S,C), however this representation is in general not unique. We show that for any CWS code Q with certain permutation symmetry, one can always find a self-dual additive code S with the same permutation symmetry as Q such that Q=(S,C). As many good CWS codes have been found by starting from a chosen S, this ensures that when trying to find CWS codes with certain permutation symmetry, the choice of S with the same symmetry will suffice. A key step for this result is a new canonical representation for CWS codes, which is given in terms of a unique decomposition as union stabilizer codes. For CWS codes, so far mainly the standard form (G,C) has been considered, where G is a graph state. We analyze the symmetry of the corresponding graph of G, which in general cannot possess the same permutation symmetry as Q. We show that it is indeed the case for the toric code on a square lattice with translational symmetry, even if its encoding graph can be chosen to be translational invariant.
Cite as

Salman Beigi, Jianxin Chen, Markus Grassl, Zhengfeng Ji, Qiang Wang, and Bei Zeng. Symmetries of Codeword Stabilized Quantum Codes. In 8th Conference on the Theory of Quantum Computation, Communication and Cryptography (TQC 2013). Leibniz International Proceedings in Informatics (LIPIcs), Volume 22, pp. 192-206, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2013)

Copy BibTex To Clipboard

@InProceedings{beigi_et_al:LIPIcs.TQC.2013.192,
  author =	{Beigi, Salman and Chen, Jianxin and Grassl, Markus and Ji, Zhengfeng and Wang, Qiang and Zeng, Bei},
  title =	{{Symmetries of Codeword Stabilized Quantum Codes}},
  booktitle =	{8th Conference on the Theory of Quantum Computation, Communication and Cryptography (TQC 2013)},
  pages =	{192--206},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-939897-55-2},
  ISSN =	{1868-8969},
  year =	{2013},
  volume =	{22},
  editor =	{Severini, Simone and Brandao, Fernando},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.TQC.2013.192},
  URN =		{urn:nbn:de:0030-drops-43129},
  doi =		{10.4230/LIPIcs.TQC.2013.192},
  annote =	{Keywords: CWS Codes, Union Stabilizer Codes, Permutation Symmetry, Toric Code}
}
  • Refine by Author
  • 2 Suk, Andrew
  • 2 Zeng, Ji
  • 1 Beigi, Salman
  • 1 Chen, Jianxin
  • 1 Grassl, Markus
  • Show More...

  • Refine by Classification
  • 2 Mathematics of computing → Combinatorics
  • 1 Computing methodologies → Neural networks

  • Refine by Keyword
  • 1 CWS Codes
  • 1 Erdős-Szekeres
  • 1 Permutation Symmetry
  • 1 Toric Code
  • 1 Union Stabilizer Codes
  • Show More...

  • Refine by Type
  • 4 document

  • Refine by Publication Year
  • 2 2022
  • 1 2013
  • 1 2023

Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail
© 2023 Schloss Dagstuhl - LZI GmbH Imprint Privacy Contact