Emmanuel Iarussi

Short Bio I am a Computer Scientist at the National Scientific and Technical Research Council (CONICET) and an Assistant Professor at the Universidad Torcuato Di Tella in Buenos Aires. I completed my Ph.D. in Computer Graphics at the GRAPHDECO team, INRIA Sophia Antipolis, under the supervision of Adrien Bousseau and George Drettakis. After graduating, I was a postdoctoral researcher in Bernd Bickel’s group at IST Austria. I also hold an Engineering degree from UNICEN University (2012). My research bridges AI, computer graphics, and medical imaging, focusing on generative models for 2D/3D content manipulation.

News

JUN 2025

Our paper VesselGPT: Autoregressive Modeling of Vascular Geometry has been accepted to MICCAI 2025. More info here.

JUN 2025

Our paper Recursive Variational Autoencoders for 3D Blood Vessel Generative Modeling has been accepted to Medical Image Analysis (Elsevier) 2025. More info here.

MAY 2025

Our paper Enhancing and Advancements in Deep Learning for Melanoma Detection: A Comprehensive Review has been published in Computers in Biology & Medicine 2025. More info here.

APR 2025

MAR 2025

I was selected to attend Khipu 2025, the premier Latin-American Machine Learning and AI conference. More info here.

FEB 2025

I participated in the workshop Exploring Opportunities for Large Language Models in Public Discourse held at UC Berkeley. More info here.

DEC 2024

Our project VisDecode: AI-Driven Interpretation and Enhancement of Scientific Plots has been awarded a $69,960 grant from the Alfred P. Sloan Foundation’s Trust in AI initiative as part of the Sloan–CZI “Pathways to AI-Enabled Research” program. More info here.

JUN 2024

I participated in the CZI Open Science 2024 Meeting held in Boston. More info here.

SEP 2024

I participated in the workshop Open Science Dynamic Convergence Workshop held in Washington, DC. More info here.

FEB 2024

Our paper DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling has been accepted to CVPR 2024. More info here.

Science communication

Publications

VesselGPT: Autoregressive Modeling of Vascular Geometry Accepted to MICCAI, 2025.

Paula Feldman, Martin Sinnona, Viviana Siless, Claudio Delrieux, Emmanuel Iarussi

Inspired by large-language-model advances, we encode vascular trees with a VQ-VAE and generate them autoregressively using a GPT-2 backbone. The model captures complex branching and geometry, enabling realistic synthesis of anatomical trees for simulation and planning […]

Recursive Variational Autoencoders for 3D Blood Vessel Generative Modeling To appear in Medical Image Analysis, 2025.

Paula Feldman, Miguel Fainstein, Viviana Siless, Claudio Delrieux, Emmanuel Iarussi

Anatomical trees play a pivotal role in clinical diagnosis and treatment planning, yet their intricate topology makes them hard to model. We introduce a recursive variational neural network that learns a compact manifold capturing branch connectivity and geometry, enabling realistic vascular synthesis […]

Enhancing and Advancements in Deep Learning for Melanoma Detection: A Comprehensive Review Computers in Biology & Medicine, 2025.

Graziela Sória Virgens, João A. Teodoro, Emmanuel Iarussi, Tiago Rodrigues, Danilo T. Amaral

This systematic review surveys recent deep-learning approaches to melanoma detection, highlighting trends, replication issues, and generalization gaps. We analyze datasets, evaluation protocols, and state-of-the-art architectures to outline future directions for reliable skin-cancer screening […]

Improving Realism in Abdominal Ultrasound Simulation Combining a Segmentation-Guided Loss and Polar Coordinates Training Medical Physics, 2025.

Santiago Vitale, José Ignacio Orlando, Emmanuel Iarussi, Alejandro Díaz, Ignacio Larrabide

We present a framework that fuses physics-based ultrasound simulation with GAN-based image translation, guided by segmentation loss and polar-coordinate training. The method reduces hallucinations and enhances anatomical fidelity, yielding realistic abdominal US images for medical training […]

DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling Accepted to CVPR, 2024.

Miguel Fainstein, Viviana Siless, Emmanuel Iarussi

In recent years, there has been a growing interest in training Neural Networks to approximate Unsigned Distance Fields (UDFs) for representing open surfaces in the context of 3D reconstruction.However, UDFs are non-differentiable at the zero level set which leads to significant errors in distances and gradients, generally resulting in [...]

VesselVAE: Recursive Variational Autoencoders for 3D Blood Vessel Synthesis International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) Springer, 2023.

Paula Feldman, Miguel Fainstein, Viviana Siless, Claudio Delrieux, Emmanuel Iarussi

We present a data-driven generative framework for synthesizing blood vessel 3D geometry. This is a challenging task due to the complexity of vascular systems, which are highly variating in shape, size, and structure. Existing model-based methods provide some degree of control and variation in the structures produced [...]

Learning normal asymmetry representations for homologous brain structures International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) Springer, 2023.

Duilio Deangeli, Emmanuel Iarussi, Juan Pablo Princich, Mariana Bendersky, Ignacio Larrabide, José Ignacio Orlando

Although normal homologous brain structures are approximately symmetrical by definition, they also have shape differences due to e.g. natural ageing. On the other hand, neurodegenerative conditions induce their own changes in this asymmetry, making them more pronounced [...]

Bone-GAN: Generation of Virtual Bone Microstructure of High Resolution Peripheral Quantitative Computed Tomography
Medical Physics American Association of Physicists in Medicine, Wiley, 2023

Felix S. L. Thomsen, Emmanuel Iarussi, Jan Borggrefe, Steven K. Boyd, Yue Wang, Michele C. Battié

This study aims to provide a reliable method for the generation of realistic bone microstructure, serving for the training of neural networks and the development of new diagnostic parameters of bone architecture and mineralization. In a first step, we trained a volumetric generative model in a progressive manner to create patches of realistic bone [...]

NORHA: A NORmal Hippocampal Asymmetry Deviation Index Based on One-Class Novelty Detection and 3D Shape Features
Brain Topography A Journal of Cerebral Function and Dynamics, Springer, 2023

Duilio Deangeli, Francisco Iarussi, Hernán Külsgaard, Delfina Braggio, Juan Pablo Princich, Mariana Bendersky, Emmanuel Iarussi, Ignacio Larrabide, José Ignacio Orlando

Radiologists routinely analyze hippocampal asymmetries in magnetic resonance (MR) images as a biomarker for neurodegenerative conditions like epilepsy and Alzheimer’s Disease. However, current clinical tools rely on either subjective evaluations, basic volume measurements, or disease-specific models that fail to capture more complex [...]

Learning deep features for dead and living breast cancer cell classification without staining Nature Scientific Reports, 2021

Gisela Pattarone, Laura Acion, Marina Simian, Emmanuel Iarussi

Automated cell classification in cancer biology is a challenging topic in computer vision and machine learning research. Breast cancer is the most common malignancy in women that usually involves phenotypically diverse populations of breast cancer cells and an heterogeneous stroma. In recent years, automated microscopy technologies are allowing the study of live cells over extended periods of time [...]

SketchZooms: Deep multi-view descriptors for matching line drawings Computer Graphics Forum Wiley, 2021

José Pablo Navarro, José Ignacio Orlando, Claudio Delrieux, Emmanuel Iarussi

Finding point-wise correspondences between images is a long-standing problem in computer vision. Corresponding sketch images is particularly challenging due to the varying nature of human style, projection distortions and viewport changes. In this paper we present a feature descriptor targeting line drawings [...]

Generative Modelling of 3D in-silico Spongiosa with Controllable Micro-Structural Parameters International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) (pp. 785-794). Springer, Cham, 2020.

Emmanuel Iarussi, Felix Thomsen, Claudio Delrieux

Research in vertebral bonemicro-structure generally requires costly procedures to obtain physical scans of real bone with a pathology under study, since no methods are available yet to generate realistic bone structures in-silico. Here we propose to apply recent advances in generative adversarial networks (GANs) to develop such a method. We adapted style-transfer techniques [...]

Improving realism in patient-specific abdominal Ultrasound simulation using CycleGANs International Journal of Computer Assisted Radiology and Surgery (pp. 1-10), 2019

Santiago Vitale, José Ignacio Orlando, Emmanuel Iarussi, Ignacio Larrabide,

In this paper we propose to apply generative adversarial neural networks trained with a cycle-consistency loss, or CycleGANs, to improve realism in ultrasound (US) simulation from Computed Tomography (CT) scans. A ray-casting US simulation approach is used to generate intermediate synthetic images [...]

FlexMaps: Computational Design of Flat Flexible Shells for Shaping 3D Objects SIGGRAPH ASIA 2018 ACM Transactions on Graphics (TOG)

Luigi Malomo, Jesús Pérez, Emmanuel Iarussi, Nico Pietroni, Eder Miguel, Paolo Cignoni, Bernd Bickel

We propose FlexMaps, a novel framework for fabricating smooth shapes out of flat, flexible panels with tailored mechanical properties. We start by mapping the 3D surface onto a 2D domain as in traditional UV mapping to design a set of deformable flat panels called FlexMaps. For these panels [...]

CoreCavity: Interactive Shell Decomposition for Fabrication with Two-Piece Rigid Molds SIGGRAPH 2018 ACM Transactions on Graphics (TOG)

Kazutaka Nakashima , Thomas Auzinger, Emmanuel Iarussi, Ran Zhang, Takeo Igarashi, Bernd Bickel

Molding is a popular mass production method, in which the initial expenses for the mold are offset by the low per-unit production cost. However, the physical fabrication constraints of the molding technique commonly restrict the shape of moldable objects. For a complex shape, a decomposition of the object into moldable parts is a common strategy [...]

WrapIt: Computer-Assisted Crafting of Wire Wrapped Jewelry SIGGRAPH ASIA 2015 ACM Transactions on Graphics (TOG)

Emmanuel Iarussi, Wilmot Li, Adrien Bousseau

Wire wrapping is a traditional form of handmade jewelry that involves bending metal wire to create intricate shapes. The technique appeals to novices and casual crafters because of its low cost, accessibility and unique aesthetic. We present a computational design tool that addresses the two main challenges of creating 2D wirewrapped [...]

BendFields: Regularized Curvature Fields from Rough Concept Sketches SIGGRAPH 2015 ACM Transactions on Graphics (TOG)

Emmanuel Iarussi, David Bommes, Adrien Bousseau

Designers frequently draw curvature lines to convey bending of smooth surfaces in concept sketches. We present a method to extrapolate curvature lines in a rough concept sketch, recovering the intended 3D curvature field and surface normal at each pixel of the sketch. This 3D information allows us to enrich the sketch with 3D-looking shading [...]

The Drawing Assistant: Automated Drawing Guidance and Feedback from Photographs ACM UIST

Emmanuel Iarussi, Adrien Bousseau, Theophanis Tsandilas

We present an interactive drawing tool that provides automated guidance over model photographs to help people practice traditional drawing-by-observation techniques. The drawing literature describes a number of techniques to support this task and help people gain consciousness of the shapes in a scene and their relationships. We compile these techniques and derive a set of construction lines that we automatically extract from a model [...]

Computer Drawing Tools for Assisting Learners, Hobbyists, and Professionals PhD. Thesis - Université de Nice

Emmanuel Iarussi

Drawing is the earliest form of visual depiction. The goal of this thesis is to facilitate and accelerate drawing for amateurs as well as for expert designers and illustrators, employing computer graphics, image processing and interaction techniques. As this is a broad spectrum to tackle, we identify three specific problems related to drawing and propose computer tools to help users overcome the main challenges on each domain [...]