Hanz Cuevas Velasquez

Research Scientist @ MPI

I am a Research Scientist at the Max Planck Institute for Intelligent Systems in Tübingen, working in Michael Black's group. My research focuses on the intersection of digital humans, generative models, and machine learning.

Previously, I obtained my PhD in 3D scene understanding at The University of Edinburgh under Prof. Bob Fisher, and worked at Microsoft Mixed Reality under Tadas Baltrušaitis.

Selected Publications

Generative Zoo
Generative Zoo: Scalable Generation of Realistic 3D Animal Pose and Shape
Tomasz Niewiadomski, Anastasios Yiannakidis, Hanz Cuevas-Velasquez, Soubhik Sanyal, Michael J. Black, Silvia Zuffi, Peter Kulits
ICCV 2025

We propose a pipeline for the scalable generation of realistic 3D animal pose and shape estimation training data.

SimpleEgo
SimpleEgo: Predicting Probabilistic Body Pose from Egocentric Cameras
Hanz Cuevas-Velasquez, Charlie Hewitt, Sadegh Aliakbarian, Tadas Baltrušaitis
3DV 2024

Predicting probabilistic body pose from egocentric cameras, addressing occlusion and limited field of view.

Gen-B
Toward Human Understanding with Controllable Synthesis
Hanz Cuevas-Velasquez, Priyanka Patel, Haiwen Feng, Michael Black
Preprint 2024

Controllable synthesis method that balances image realism with precise ground truth (Generative BEDLAM).

Novel Head Pose
Novel Head Pose, Expression, and Gaze Synthesis
Hanz Cuevas-Velasquez
Project

Synthesis of novel head poses, expressions, and gazes from short cellphone videos.

Procedural Humans
Procedural Humans for Computer Vision
Charlie Hewitt, Tadas Baltrušaitis, Erroll Wood, Lohit Petikam, Louis Florentin, Hanz Cuevas Velasquez
Microsoft Research Report

Parametric model of the face and body and rendering pipeline to generate realistic images of humans.

Ge-Latto
Two Heads are Better than One: Geometric-Latent Attention
Hanz Cuevas-Velasquez, Antonio Javier Gallego, Robert B. Fisher
BMVC 2021

Two-headed attention layer combining geometric and latent features for point cloud segmentation.

Trimbot
Real-time Stereo Visual Servoing for Rose Pruning
Hanz Cuevas-Velasquez, et al.
ICRA 2020

Visual servoing to find and cut rose branches in real-time using a robotic arm.

Short CV

2023 - Present

Research Scientist

Max Planck Institute - Tuebingen (Michael Black's group)

2023

Visiting Researcher

The University of Edinburgh

2022

Researcher Intern

Microsoft

2017 - 2022

Ph.D. in Informatics

The University of Edinburgh

2016 - 2017

MSc. Artificial Intelligence

The University of Edinburgh

2011 - 2015

BSc. Mechatronics Engineering

Universidad Catolica Boliviana "San Pablo"

Visitor Map