If like me you’re using **Beamer** to create slides in LaTeX, there’s a neat trick that helps you include SVG vector graphics—and even better, animate them progressively using Inkscape layers and Beamer’s overlay features like `\pause` or `\onslide`. ## Why SVG and Inkscape? SVG (Scalable Vector Graphics) is a great format for high-quality images that scale well without losing resolution. If you create your SVGs using **Inkscape** , you can organize your graphic into **layers**. These layers can then be revealed step-by-step in your presentation using Beamer’s overlay commands. This avoids the hassle of creating multiple image files or slides to simulate animation, and it keeps everything clean and well aligned on all slides displaying your image. ## SVG image inclusion in LaTeX To include SVGs in LaTeX and make use of LaTeX commands inside the image (like math formulas or labels), you can use the svg package. This package allows you to embed SVG files and even interpret LaTeX code within them during compilation. Here’s a typical command to include an SVG: \usepackage{svg} ... \includesvg[inkscapelatex=true, width=\textwidth]{<image.svg>} It’s thus very similar to the command `\includegraphics` of the package `graphicx` (see here). For more information, you can have a look to the full description of the SVG package on CTAN (check there the full documentation in pdf). ## Step-by-Step Animation with Layers In Inkscape, you can create multiple layers and insert LaTeX overlay commands like `\pause` or `\onslide` directly into the layer content. When compiled with Beamer, these layers will appear progressively, creating a smooth animation effect. There’s for instance this video tutorial showing how to set up layers in Inkscape for this purpose. ## What You Need To make this work, you’ll need: * **Inkscape** installed on your system (check here for installation instruction) * LaTeX compilation with the `-shell-escape` option enabled (this allows LaTeX to call Inkscape during compilation). For instance, if you use `pdflatex` the command to run is pdflatex -shell-escape <latexfilename> _Remark_ : The `-shell-escape` option should be used with caution in general as this gives the possibility to latex compilation to execute shell commands. ## Real-World Use I use this technique regularly in my course and seminar slides. It’s a great way to present complex diagrams step-by-step without cluttering your slides or manually managing multiple image versions. Here is a minimal example as well as this minimal SVG file you can use (note that if you visualize it in your browser by clicking on the link, you’ll see all the layers at once, as well as the `\pause` command). Note that this example must be compiled with the command `pdflatex -shell-escape MiniEx.tex` in a directory including the SVG file `test.svg`. All the sources of this example are included in this zip file. The produced minimal PDF presentation is available here. %% File: MiniEx.tex \documentclass{beamer} \title{Minimal Beamer Example with Inscape Overlays} \usepackage{graphicx} \usepackage{svg} \begin{document} \frame{\titlepage} \begin{frame} \begin{center} \includesvg[inkscapelatex=true,width=.4\textwidth]{test.svg} \end{center} \end{frame} \end{document}

**Abstract** : Representing and processing data in spherical domains presents unique challenges, primarily due to the curvature of the domain, which complicates the application of classical Euclidean techniques. Implicit neural representations (INRs) have emerged as a promising alternative for high-fidelity data representation; however, to effectively handle spherical domains, these methods must be adapted to the inherent geometry of the sphere to maintain both accuracy and stability. In this context, we propose Herglotz-NET (HNET), a novel INR architecture that employs a harmonic positional encoding based on complex Herglotz mappings. This encoding yields a well-posed representation on the sphere with interpretable and robust spectral properties. Moreover, we present a unified expressivity analysis showing that any spherical-based INR satisfying a mild condition exhibits a predictable spectral expansion that scales with network depth. Our results establish HNET as a scalable and flexible framework for accurate modeling of spherical data.

**Abstract** : In this paper, we tackle the issue of implementing a dithering procedure for the 1-bit quantization of radar signals that is able to generate high-quality estimates while remaining a low-complexity and cost-efficient solution. Classic Dithering techniques that add a signal before the Analog to Digital Converter (ADC) have been used in many acquisition chain designs and have been studied as a way to shape the quantization noise more favourably. Here, we stray away from this additive dithering, which, as will be made clear later, induces a complex and high-cost implementation. Instead, we propose the use of a multiplicative phase dithering. This process can leverage already existing radar architectures of Frequency Modulated Continuous Wave (FMCW) radars and can thus be efficiently implemented. The efficiency of this multiplicative dithering is first studied theoretically, and its link to another coarse quantization scheme, namely the Phase-Only acquisition, is highlighted. The performances of this novel dithering scheme are then extensively tested using Monte Carlo simulations and are thoroughly compared to their additive counterparts. A hardware-relaxed version of the random phase dithering is also introduced and compared to the other 1-bit schemes. The observations made in simulations are then validated using actual radar measurements at 24 GHz. Combined with the simulations, these measurements show that the multiplicative dithering is an appealing alternative to the additive random dithering in a low number of measurement setting. Specifically, we show that this procedure is a good trade-off between strong theoretical guarantees and reconstruction quality for low-complexity hardware.