Few-Shot Learning with a Novel Voronoi Tessellation-Based Image Augmentation Method for Facial Palsy Detection

Face palsy has adverse effects on the appearance of a person and has negative social and functional consequences on the patient. Deep learning methods can improve face palsy detection rate, but their efficiency is limited by insufficient data, class imbalance, and high misclassification rate. To alleviate the lack of data and improve the performance of deep learning models for palsy face detection, data augmentation methods can be used. In this paper, we propose a novel Voronoi decomposition-based random region erasing (VDRRE) image augmentation method consisting of partitioning images into randomly defined Voronoi cells as an alternative to rectangular based random erasing method. The proposed method augments the image dataset with new images, which are used to train the deep neural network. We achieved an accuracy of 99.34% using two-shot learning with VDRRE augmentation on palsy faces from Youtube Face Palsy (YFP) dataset, while normal faces are taken from Caltech Face Database. Our model shows an improvement over state-of-the-art methods in the detection of facial palsy from a small dataset of face images.

Generative Infills for Additive Manufacturing Using Space-Filling Polygonal Tiles

We study a new class of infill patterns, that we call wallpaper-infills for additive manufacturing based on space-filling shapes. To this end, we present a simple yet powerful geometric modeling framework that combines the idea of Voronoi decomposition space with wallpaper symmetries defined in 2-space. We first provide a geometric algorithm to generate wallpaper-infills and design four special cases based on selective spatial arrangement of seed points on the plane. Second, we provide a relationship between the infill percentage to the spatial resolution of the seed points for our cases thus allowing for a systematic way to generate infills at the desired volumetric infill percentages. Finally, we conduct a detailed experimental evaluation of the of these four cases to study their mechanical behavior under tensile loading.

Voronoi Decomposition of Cardiovascular Dependency Structures in Different Ambient Conditions: An Entropy Study

This paper proposes a method that maps the coupling strength of an arbitrary number of signals D, D ≥ 2, into a single time series. It is motivated by the inability of multiscale entropy to jointly analyze more than two signals. The coupling strength is determined using the copula density defined over a [0 1]D copula domain. The copula domain is decomposed into the Voronoi regions, with volumes inversely proportional to the dependency level (coupling strength) of the observed joint signals. A stream of dependency levels, ordered in time, creates a new time series that shows the fluctuation of the signals’ coupling strength along the time axis. The composite multiscale entropy (CMSE) is then applied to three signals, systolic blood pressure (SBP), pulse interval (PI), and body temperature (tB), simultaneously recorded from rats exposed to different ambient temperatures (tA). The obtained results are consistent with the results from the classical studies, and the method itself offers more levels of freedom than the classical analysis.

SEAMLESS IMAGE MOSAICKING VIA SYNCHRONIZATION

This paper proposes an innovative method to create high-quality seamless planar mosaics. The developed pipeline ensures good robustness against many common mosaicking problems (e.g., misalignments, colour distortion, moving objects, parallax) and differs from other works in the literature because a global approach, known as <i>synchronization</i>, is used for image registration and colour correction. To better conceal the mosaic seamlines, images are cut along specific paths, computed using a Voronoi decomposition of the mosaic area and a shortest path algorithm. Results obtained on challenging real datasets show that the colour correction mitigates significantly the colour variations between the original images and the seams on the final mosaic are not evident.

Multimodal Information Systems

The number of individuals using public transportation is increasing. Transport companies want to ensure, at best, the satisfaction of the travellers. Nevertheless, a significant number of these companies sometimes pushes the travellers to confusion to compose their itineraries and obtain the required information. The authors suggest in this chapter integrating several traveller information systems into the same global system. This chapter aims to provide information to the traveller without concern for their location and optimize processing by limiting the number of involved nodes. They opted for a multi-agent system associated with the Voronoï decomposition of the global network.

A Numerical Characterization of the Nanoparticles Distribution on the Surface of a Semiconductor

The motivation for this work was to qualitatively describe the distribution of Au nanoparticles on the surface of a semiconductor. We discuss suitable mathematical characteristics which allow the uniform distribution to be distinguished from the distribution a ected by any physical phenomenon, i.e. by the repulsive force between electrically charged particles or by the influence of properties of the surface. We identify Voronoi decomposition and a statistical analysis of Voronoi cell properties as a suitable tool for this purpose.