Felipe Torres Figueroa

Computer Vision & Deep Learning Research Scientist

Home

I'm interested in Machine Learning and Computer Vision (CV) regarding new technologies or applications. Currently, I'm working on explainable AI in image recognition. I'm interested in foundational models, as well as Large Language Models (LLMs), and the overall state of the art such as the transformer family of methods.
I recently finished my PhD on Interpretable Image Recognition, under the supervision of Ronan Sicre, Stéphane Ayache and Yannis Avrithis. The supporting material for my defense can be found in my thesis slides, and the support material.
On September 8th 2025 I joined Institute Mines-Telecom Business School to contribute to the ENFIELD project with the objective of developing Interpretable and Trustworthy solutions to key areas such as Healthcare and Space. I to expect to bring significative contributions to these domains given my prior experience and current abilities, expanding the understanding of existing AI methods.

ftorres11 felitf.94@gmail.com LinkedIn ORCID Scholar

Research

Saliency Maps Give a False Sense of Explanability to Image Classifiers: An Empirical Evaluation across Methods and Metrics
H. Zhang, F. Torres, H. Hermanns.
Asian Conference on Machine Learning, 2024.
PDF BibTex

CA-Stream: Attention-based pooling for interpretable image recognition
F. Torres, H. Zhang, R. Sicre, S. Ayache, Y. Avrithis.
CVPR Workshop on Explainable AI, 2024.
PDF BibTex

A Learning Paradigm for Interpretable Gradients
F. Torres, H. Zhang, R. Sicre, Y. Avrithis, S. Ayache.
VISAPP 2024. Oral.
PDF BibTex

Opti-CAM: Optimizing saliency maps for interpretability
H. Zhang, F. Torres, R. Sicre, Y. Avrithis, S. Ayache.
ArXiv, 2023.
PDF BibTex

Countor: count without bells and whistles
A. Ospina, F. Torres.
CVPR Workshop, 2020.
PDF BibTex

An ImageJ plugin for the high throughput image analysis of in vitro scratch wound healing assays
A. Suarez, F. Torres, C. Clavijo, P. Arbeláez, J. Cruz, C. Muñoz.
PLOS ONE, 2020.
PDF BibTex

Hand Pose Estimation for Pediatric Bone Age Assessment
M. C. Escobar, C. I. González, F. Torres, L. Daza, G. Triana and P. Arbeláez.
MICCAI-2019 Oral.
PDF BibTex Project Page

About

I'm 30 years old, passionate about music, history and astronomy. During my undergraduate I grew passion towards programming, and since then I've been working on Computer Vision and Deep Learning. I used to be good at jogging until I took a mountain to the (left) knee. I love spending time looking at the sky, languages and listening to music; specially if I can reproduce it when playing my guitar or bass. Whenever I have the means to, I also build scale models of planes of the WWII era. Ever since the second part of 2021 I picked up an appreciation to Asian cuisine, specially Chinese spicy dishes from the Guizhou region.