arXiv research

Proposes a fixed smooth convolutional layer to reduce checkerboard artifacts in CNNs.

We analyze architectural features of Deep Neural Networks (DNNs) using the so-called Neural Tangent Kernel (NTK), which describes the training and generalization of DNNs in the infinite-width setting. In this setting, we show that for fully-connected DNNs, as the depth grows, two regimes appear: "order", where the (sca…

Paper detects checkerboard colorability of virtual links using odd writhe and arrow polynomial.

New polynomial for checkerboard-colorable 4-valent virtual graphs.

The paper confirms a conjecture and extends arrow polynomial to twisted links.

Proves certain alternating links have specific geometric properties.

A new copula, the checkerboard copula, maximizes entropy and preserves dependence.

Geometric duality connects graph isomorphism and knot equivalence.

Associated to every state surface for a knot or link is a state graph, which embeds as a spine of the state surface. A state graph can be decomposed along cut-vertices into graphs with induced planar embeddings. Associated with each such planar graph is a checkerboard surface, and each state surface is a fiber if and o…

Checkerboard surfaces in alternating link complements are used frequently to determine information about the link. However, when many crossings are added to a single twist region of a link diagram, the geometry of the link complement stabilizes (approaches a geometric limit), but a corresponding checkerboard surface in…

We define an equivalence relation on graphs with signed edges, such that the associated adjacency matrices of two equivalent graphs are congruent over $\mathbb{Z}$. We show that signed graphs whose eigenvalues are larger than $-2$ are equivalent to one of the simply laced Dynkin diagrams: $A_{n}$, $D_{n}$, $E_{6}$, $E_…

The notion of chckerboard colorability for virtual links and abstract links is introduced. We study the Jones polynomials of virtual links and abstruct links. It is proved that a certain property of the Jones polynomials of classical links is valid for virtual links which admit checkerboard colorings.

We associate an open book with any connected plane checkerboard graph, thus providing a common extension of the classes of prime positive braid links and positive tree-like Hopf plumbings. As an application, we prove that the link type of a prime positive braid closure is determined by the linking graph associated with…

The paper explores discrete isothermic nets using checkerboard patterns in quadrilateral nets.

The paper studies right-angled links on higher genus surfaces.

We present a procedure which allows one to integrate explicitly the class of checkerboard IC-nets which has recently been introduced as a generalisation of incircular (IC) nets. The latter class of privileged congruences of lines in the plane is known to admit a great variety of geometric properties which are also pres…

Determinant modulo 8 classifies virtual knots based on polynomial coefficients.

We consider congruences of straight lines in a plane with the combinatorics of the square grid, with all elementary quadrilaterals possessing an incircle. It is shown that all the vertices of such nets (we call them incircular or IC-nets) lie on confocal conics. Our main new results are on checkerboard IC-nets in the p…

Extends Gordon-Litherland pairing to links in thickened surfaces, defining new invariants.

Essential surfaces in link diagrams on surfaces are crucial for understanding link properties.

Single CNN removes multiple ultrasound artifacts.

Study examines how image artifacts impact polyp detection and proposes methods to mitigate their effects.

In this paper, we develop a convolutional neural network model to predict the mechanical properties of a two-dimensional checkerboard composite quantitatively. The checkerboard composite possesses two phases, one phase is soft and ductile while the other is stiff and brittle. The ground-truth data used in the training …

New method corrects motion artifacts in MR images without paired data.

Electroencephalograms (EEG) are often contaminated by artifacts which make interpreting them more challenging for clinicians. Hence, automated artifact recognition systems have the potential to aid the clinical workflow. In this abstract, we share the first results on applying various machine learning algorithms to the…

New invariant for virtual links defined using homology.

Characterizes arithmetic and commensurable links in curved surfaces.

It is shown that there exist alternating non-Montesinos knots whose essential spanning surfaces with maximal and minimal boundary slopes are not realised by the checkerboard surfaces coming from a reduced alternating planar diagram.

Voice conversion (VC) aims at conversion of speaker characteristic without altering content. Due to training data limitations and modeling imperfections, it is difficult to achieve believable speaker mimicry without introducing processing artifacts; performance assessment of VC, therefore, usually involves both speaker…

This monograph derives direct and concrete relations between colored Jones polynomials and the topology of incompressible spanning surfaces in knot and link complements. Under mild diagrammatic hypotheses that arise naturally in the study of knot polynomial invariants (A- or B-adequacy), we prove that the growth of the…

In this paper we review the definitions of homogeneous and alternative links. We also give two new characterizations of an alternative link diagram, one within the context of the enhanced checkerboard graph and another from the labeled Seifert graph.

Study geodesic X-ray transform and streaking artifacts on simple surfaces or spaces of constant curvature.

In this paper, we study the Khovanov homology of an alternating virtual link $L$ and show that it is supported on $g+2$ diagonal lines, where $g$ equals the virtual genus of $L$. Specifically, we show that $Kh^{i,j}(L)$ is supported on the lines $j=2i-σ_ξ+2k-1$ for $0\leq k\leq g+1$ where $σ_{ξ^*}(L)+2g= σ_ξ(L)$ are th…

Paper introduces adversarial lossy compression for video artifacts reduction.

New spanning tree model connects knot homology, s-invariant, and exotic discs.

A new unsupervised method removes CT metal artifacts using beta-CycleGAN and attention.

Single model corrects JPEG artifacts for various compression settings.

Accelerated magnetic resonance (MR) scan acquisition with compressed sensing (CS) and parallel imaging is a powerful method to reduce MR imaging scan time. However, many reconstruction algorithms have high computational costs. To address this, we investigate deep residual learning networks to remove aliasing artifacts …

Improved linear upper bound for ribbonlength of knots.

We show that the Kauffman bracket $[L]$ of a checkerboard colorable virtual link $L$ is an evaluation of the Bollobás-Riordan polynomial $R_{G_L}$ of a ribbon graph associated with $L$. This result generalizes Thistlethwaite's celebrated theorem relating the Kauffman bracket with the Tutte polynomial of planar graphs.

Paper explores unsupervised learning for ultrasound image artifact removal.

A group-theoretical method, via Wada's representations, is presented to distinguish Kishino's virtual knot from the unknot. Biquandles are constructed for any group using Wada's braid group representations. Cocycle invariants for these biquandles are studied. These invariants are applied to show the non-existence of Al…

Conebeam CT using a circular trajectory is quite often used for various applications due to its relative simple geometry. For conebeam geometry, Feldkamp, Davis and Kress algorithm is regarded as the standard reconstruction method, but this algorithm suffers from so-called conebeam artifacts as the cone angle increases…

Interpretation of electroencephalogram (EEG) signals can be complicated by obfuscating artifacts. Artifact detection plays an important role in the observation and analysis of EEG signals. Spatial information contained in the placement of the electrodes can be exploited to accurately detect artifacts. However, when few…

L-CNNs approximate gauge actions, revealing fixed points with no lattice artifacts.

Generative model predicts menstrual cycle lengths accounting for self-tracking artifacts.

Estimates support in distributions with sampling artifacts and errors.

Deconvolutional layers have been widely used in a variety of deep models for up-sampling, including encoder-decoder networks for semantic segmentation and deep generative models for unsupervised learning. One of the key limitations of deconvolutional operations is that they result in the so-called checkerboard problem.…