Cristian Canton

I am the head of Responsible AI (RAI) at Meta, where I support multiple organizations related to AI robustness, fairness, transparency, and legitimacy. Recently, have expanded my scope to cover all aspects of RAI for Generative AI.

Erdős number: 3.

LinkedIn /  Threads /  Twitter /  Google Scholar /  Email

profile photo
Research

I'm interested in computer vision, machine learning, security in AI, adversarial attacks and fairness.

Towards Red Teaming in Multimodal and Multilingual Translation
Christophe Ropers, David Dale, Prangthip Hansanti, Gabriel Mejia Gonzalez, Ivan Evtimov, Corinne Wong, Christophe Touret, Kristina Pereyra, Seohyun Sonia Kim, Cristian Canton Ferrer, Pierre Andrews, Marta R. Costa-jussà
Submitted to "BIG CONFERENCE", 2024.
[bibtex]

TL;DR. This paper presents the first study on human based red teaming for Machine Translation,marking a significant step towards understanding and improving the performance of translation models.

Data-Driven but Privacy-Conscious: Pedestrian Dataset De-identification via Full-Body Person Synthesis
Maxim Maximov, Tim Meinhardt, Ismail Elezi, Zoe Papakipos, Caner Hazirbas, Cristian Canton Ferrer, Laura Leal-Taixé
IEEE International Conference on Automatic Face and Gesture Recognition, Istanbul (Turkey), 2024.
[bibtex]

TL;DR. In this paper we prove that the pedestrian tracking task can be done in a privacy preserving manner by replacing individuals by time-consistent full-body person synthetic.

Code Llama: Open Foundation Models for Code
Baptiste Rozière, Jonas Gehring, ..., Cristian Canton Ferrer, and many more amazing authors.
Meta AI Technical Report, 2023.
[bibtex]  [GitHub]  [Model]

TL;DR. Code Llama is an AI model built on top of Llama 2, fine-tuned for generating and discussing code. First of it kind!

Llama 2: Open Foundation and Fine-Tuned Chat Models
Hugo Touvron, Louis Martin, ..., Cristian Canton Ferrer, and many more amazing authors.
Meta AI Technical Report, 2023.
[bibtex]  [dataset]

TL;DR. This paper introduces Llama2, a collection of pretrained and fine-tuned large language models (LLMs) ranging in scale from 7 billion to 70 billion parameters.

On Responsible Machine Learning Datasets with Fairness, Privacy, and Regulatory Norms
Surbhi Mittal, Kartik Thakral, Richa Singh, Mayank Vatsa, Tamar Glaser, Cristian Canton Ferrer, Tal Hassner
Submitted to Nature Communications, 2023.
[bibtex]

TL;DR. A summary paper evaluating fairness, privacy and regulatory aspects of over 100 datasets. A work that sets the standards on defining what a responsible dataset is.

VPA: Fully Test-Time Visual Prompt Adaptation
Jiachen Sun, Mark Ibrahim, Melissa Hall, Ivan Evtimov, Zhuoqing Mao, Cristian Canton Ferrer, Caner Hazirbas
ACM International Conference on Multimedia (ACM MM), Ottawa (Canada), 2023.
[bibtex]

TL;DR. The first framework that generalizes visual prompting with test-time adaptation. VPA introduces a small number of learnable tokens, enabling fully test-time and storage-efficient adaptation without necessitating source-domain information.

The Casual Conversations v2 Dataset
Bilal Porgali, Vítor Albiero, Jordan Ryda, Cristian Canton Ferrer, Caner Hazirbas
Computer Vision and Pattern Recognition Conference (CVPR). Workshop on Fair, Data-efficient, and Trusted Computer Vision, Vancouver (Canada), 2023.
[bibtex]

TL;DR. A diverse, large benchmark for measuring fairness and robustness in audio/vision/speech models. A bold step forward from our Casual Conversations v1 dataset.

Designing Casual Conversations v2: A large consent-driven dataset to measure algorithmic bias and robustness
Caner Hazirbas, Yejin Bang, Tiezheng Yu, Parisa Assar, Bilal Porgali, Vítor Albiero, Stefan Hermanek, Jacqueline Pan, Emily McReynolds, Miranda Bogen, Pascale Fung, Cristian Canton Ferrer
Meta AI Report, 2023.
[bibtex]

TL;DR. A walk through on how to design a dataset that is useful to measure gaps in bias and robustness for computer vision and speech.

A Whac-A-Mole Dilemma: Shortcuts Come in Multiples Where Mitigating One Amplifies Others
Zhiheng Li, Ivan Evtimov, Albert Gordo, Caner Hazirbas, Tal Hassner, Cristian Canton Ferrer, Chenliang Xu, Mark Ibrahim
Computer Vision and Pattern Recognition Conference (CVPR), Vancouver (Canada), 2023.
[bibtex]

TL;DR. Unintended decision rules that are unable to generalize, shortcuts, are very prevalent in ML models. This paper how mitigations for shortcuts are a whack-a-model game, for now.

Affective Signals in a Social Media Recommender System
Jane Dwivedi-Yu, Yi-Chia Wang, Lijing Qin, Cristian Canton, Alon Y. Halevy
ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), Washington DC, (USA), 2022.
[bibtex]

TL;DR. A first paper to design a ranking function that gives useful and personalized post recommendations in social media.

Generating High Fidelity Data from Low-density Regions using Diffusion Models
Vikash Sehwag, Caner Hazirbas, Albert Gordo, Firat Ozgenel, Cristian Canton
Computer Vision and Pattern Recognition Conference (CVPR), New Orleans, (USA), 2022.
[bibtex]

TL;DR. Leverage diffusion process based generative models to synthesize novel images from low-density regions.

PixInWav: Residual Steganography for Hiding Pixels in Audio
Margarita Geleta, Cristina Punti, Kevin McGuinness, Jordi Pons, Cristian Canton, Xavier Giro-i-Nieto
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Singapore, 2022.
[bibtex]

TL;DR. A pioneering work on hidding images on audio, showing real results retrieving images from recorded audio waves.

Results and findings of the 2021 Image Similarity Challenge
Zoe Papakipos, Giorgos Tolias, Tomas Jenicek, Ed Pizzi, Shuhei Yokoo, Wenhao Wang, Yifan Sun, Weipu Zhang, Yi Yang, Sanjay Addicam, Sergio Manuel Papadakis, Cristian Canton, Ondrej Chum, Matthijs Douze
NeurIPS Workshop Summary, 2021.
[bibtex]

TL;DR. A quantitative and qualitative analysis of the top submissions to the 2021 Image Similarity Challenge.

The 2021 Image Similarity Dataset and Challenge
Matthijs Douze, Giorgos Tolias, Ed Pizzi, Zoë Papakipos, Lowik Chanussot, Filip Radenovic, Tomas Jenicek, Maxim Maximov, Laura Leal-Taixé, Ismail Elezi, Ondřej Chum, Cristian Canton
Facebook AI Technical Report, 2021.
[bibtex]  [Competition Track 1]  [Competition Track 2]

TL;DR. Prove how good are you at finding modified copies of images. $200K in prizes!

Adversarial Evaluation of Multimodal Models under Realistic Gray Box Assumption
Ivan Evtimov, Russ Howes, Brian Dolhansky, Hamed Firooz, Cristian Canton
Computer Vision and Pattern Recognition Conference (CVPR).
Workshop on Adversarial Machine Learning
, 2021. Best paper award
[bibtex]

TL;DR. Multimodal understanding (image+text) can be attacked in a gray-box fashing. Examples on hate memes included!

Localized Uncertainty Attacks
Ousmane Dia, Theofanis Karaletsos, Caner Hazirbas, Cristian Canton,
Ilknur Kaynar Kabul, Erik Meijer
Computer Vision and Pattern Recognition Conference (CVPR).
Workshop on Adversarial Machine Learning
, 2021.
[bibtex]

TL;DR. Undertand what areas of an image are more vulnerable to increase uncertainty of models and eventually making them to produce wrong decisions.

Adversarial collision attacks on image hashing functions
Brian Dolhansky, Cristian Canton
Computer Vision and Pattern Recognition Conference (CVPR).
Workshop on Adversarial Machine Learning
, 2021.
[bibtex]

TL;DR. Hashing algorithms (in general) can be easily compromised by introducing collisions.

Towards measuring fairness in AI: the Casual Conversations dataset
Caner Hazirbas, Joanna Bitton, Brian Dolhansky, Jacqueline Pan,
Albert Gordo, Cristian Canton
Computer Vision and Pattern Recognition Conference (CVPR).
Workshop on Responsible Computer Vision
, 2021.
IEEE Transactions on Biometrics, Behavior, and Identity Science (TBIOM), 2021.
[bibtex]  [dataset]

TL;DR. A large dataset (10TB, 30+ days of video, 3K people, 15 mins per person) featuring a diversity of actors with age, gender and skin tone annotations.

Adversarial Threats to DeepFake Detection: A Practical Perspective
Shehzeen Hussain, Paarth Neekhara, Brian Dolhansky, Joanna Bitton, Cristian Canton, Julian McAuley, Farinaz Koushanfar
Journal - ACM Digital Threats: Research and Practice, 2021.
[bibtex]

TL;DR. Defusing deepfake detection by crafting universal adversarial perturbations, transferable across different detection models.

Adversarial Threats to DeepFake Detection: A Practical Perspective
Paarth Neekhara, Brian Dolhansky, Joanna Bitton, Cristian Canton
Computer Vision and Pattern Recognition Conference (CVPR).
Workshop on Media Forensics
, 2021.
[bibtex]

TL;DR. Creation of adversarial additive patterns against open sourced deepfake detection systems turns out to be very effective.

Deep Poisoning: Towards Robust Image Data Sharing against Visual Disclosure
Hao Guo, Brian Dolhansky, Eric Hsin, Phong Dinh, Song Wang, Cristian Canton
IEEE Winter Conference on Applications of Computer Vision (WACV), 2021.
[bibtex]

TL;DR. Create embeddings that are privacy preserving thus non-reversible to obtain the original image.

Preserving Integrity in Online Social Networks
Alon Halevy, Cristian Canton, Hao Ma, Umut Ozertem, Patrick Pantel,
Marzieh Saeidi, Fabrizio Silvestri, Ves Stoyanov
Journal - Communications of the ACM (CACM), 2021.
[bibtex]

TL;DR. A comprehensive survey of AI techniques applied in social networks to preserve integrity and preserve users from bad experiences.

The Deepfake Detection Challenge Dataset
Brian Dolhansky, Joanna Bitton, Ben Pflaum, Jikuo Lu, Russ Howes, Menglin Wang,
Cristian Canton
Facebook AI Technical Report, 2020.
[bibtex]  [dataset]  [Kaggle competition]

TL;DR. Up to today, the largest deepfake dataset. A $1M competition!

The Deepfake Detection Challenge (DFDC) Preview Dataset
Brian Dolhansky, Russ Howes, Ben Pflaum, Nicole Baram, Cristian Canton
Facebook AI Technical Report, 2019.
[bibtex]  [dataset]  [blog post]

TL;DR. A small preview version of the DFDC dataset to come later.

Hate Speech in Pixels: Detection of Offensive Memes towards Automatic Moderation
Benet Oriol Sabat, Cristian Canton, Xavier Giró‑i‑Nieto
NeurIPS AI for Social Good Workshop, Vancouver, (Canada), 2019.
[bibtex]

TL;DR. Multimodal understanding of memes can allow better detection of image+text hate speech.

Towards social pattern characterization in egocentric photo-streams
Maedeh Aghaei, Mariella Dimiccoli, Cristian Canton, Petia Radeva
Journal - Computer Vision and Image Understanding (CVIU), 2018.
[bibtex]

TL;DR. Leverage ego-centric camera footage to understand long time span scene interactions, specially for social style signals..

Eye In-Painting with Exemplar Generative Adversarial Networks
Brian Dolhansky, Cristian Canton
Computer Vision and Pattern Recognition Conference (CVPR), Salt Lake City, (USA), 2018.
[bibtex]  [poster]

TL;DR. Given an example of how your eyes look, this algorithm can take a pic with your eyes closed and open them. GAN-powered.

Social Style Characterization from Egocentric Photo-Streams
Maedeh Aghaei, Mariella Dimiccoli, Cristian Canton, Petia Radeva
International Conference on Computer Vision (ICCV). Workshop on Egocentric Percetion, Interaction and Computing, Venice (Italy), 2017.
[bibtex]

TL;DR. Use an ego-camera, process multimodal inputs to do long time span scene understanding, specially for social style signals.

Disentangling Motion, Foreground and Background Features in Videos
Xunyu Lin, Víctor Campos, Xavier Giró‑i‑Nieto, Jordi Torres, Cristian Canton
Computer Vision and Pattern Recognition Conference (CVPR). Brave New Motion Representations Workshop, Honolulu (USA), 2017.
[bibtex]  [source code]  [poster]

TL;DR. An unsupervised framework to extract semantically rich features for video representation inspired by how the human visual system groups objects based on motion cues.

SalGAN: Visual Saliency Prediction with Generative Adversarial Networks
Junting Pan, Cristian Canton, Kevin McGuinness, Noel O'Connor, Jordi Torres, Elisa Sayrol, Xavier Giró‑i‑Nieto
Computer Vision and Pattern Recognition Conference (CVPR). Scene Understanding Workshop, Honolulu (USA), 2017.
[bibtex]  [source code]  [poster]

TL;DR. Estimate of saliency using a GAN, probably the first article to do it!

Training Deep Networks for Facial Expression Recognition with Crowd-Sourced Label Distribution
Emad Barsoum, Cha Zhang, Cristian Canton, Zhengyou Zhang
ACM International Conference on Multimodal Interaction (ICMI), Tokio (Japan), 2016.
[bibtex]  [dataset]

TL;DR. Some strategies to combine multiple subjective labels for human emotions in a way that CNNs can be trained effectively.

Emotion Recognition in the Wild from Videos using Images
Sarah Bargal, Emad Barsoum, Cristian Canton, Cha Zhang
ACM International Conference on Multimodal Interaction (ICMI). Workshop on Facial Expressions in the Wild, Tokio (Japan), 2016.
[bibtex]

TL;DR. Ensambles of CNN embeddings allow for high accuracy emotion recognition on videos.

Learning by Tracking: Siamese CNN for Robust Target Association
Laura Leal-Taixé, Cristian Canton, Konrad Schindler
Computer Vision and Pattern Recognition Conference (CVPR). DeepVision Workshop, Las Vegas (USA), 2016.
[bibtex]  [poster]

TL;DR. While tracking multiple targets, comparing detections using a Siamese network improves overall performance.

Motion Capture Technologies for Pose Estimation
Cristian Canton, P. Smyth
Book chapter - Skill Training in Multimodal Virtual Environments (Taylor & Francis Ed.), 2012.
[bibtex]

TL;DR. A survey on marker based motion capture technologies for pose estimation.

Multi-Camera Multi-Object Voxel Based Monte Carlo 3D Tracking Strategies
Cristian Canton, Josep R. Casas, Montse Pardàs, Enric Monte
Journal - EURASIP Journal on Advances on Signal Processing, 2011.
[bibtex]

TL;DR. Tracking multiple object on the 3D domain by voxel reconstruction and particle filtering techniques.

Human Motion Capture Using Scalable Body Models
Cristian Canton, Josep R. Casas, Montse Pardàs
Journal - Computer Vision and Image Understanding (CVIU), 2011.
[bibtex]

TL;DR. A new Monte Carlo method applied to recover articulated structured (human body) exploiting hierarchies of complexity.

Acoustic Event Detection based on Feature-Level Fusion of Audio and Video Modalities
Taras Butko, Cristian Canton, Carlos Segura, Xavier Giró-i-Nieto, Climent Nadeu,
Javier Hernando, Josep R. Casas
Journal - EURASIP Journal on Advances on Signal Processing, 2011.
[bibtex]

TL;DR. Combination of motion cues, position and audio allows detection of events within a SmartRoom environment.

Image and Video Processing Tools for HCI
Montse Pardàs, Verónica Vilaplana, Cristian Canton
Book chapter - Multimodal Signal Processing: Theory and applications for human-computer interaction (Academic Press), 2010.
[bibtex]

TL;DR. Survey on image and video processing techniques for human computer interaction in SmartRooms.

Virtual View Appearance Representation for Human Motion Analysis in Multi-View Environments
Adolfo López, Cristian Canton, Josep R. Casas,
European Signal Processing Conference (EUSIPCO), Aalborg (Denmark), 2010.
[bibtex]

TL;DR. Generating virtual orthogonal views derived from 3D voxels allow better action recognition on 2D projections.

Marker-based Human Motion Capture in Multi-View Sequences
Cristian Canton, Josep R. Casas, Montse Pardàs
Journal - EURASIP Journal on Advances on Signal Processing, 2010.
[bibtex]

TL;DR. Marker based motion tracking using annealed particle filtering.

Spatio-Temporal Alignment and Hyperspherical Radon Transform for 3D Gait Recognition in Multi-View Environments
Cristian Canton, Josep R. Casas, Montse Pardàs
IEEE International Computer Vision and Pattern Recognition Conference (CVPR). Biometrics Workshop, San Francisco (USA), 2010.
[bibtex]

TL;DR. Understanding actions directly on 3D+time hypervolumes using the Radon transform.

Real-Time 3D Multi-Person Tracking Using Monte Carlo Surface Sampling
Cristian Canton, Josep R. Casas, Montse Pardàs
IEEE International Computer Vision and Pattern Recognition Conference (CVPR). Computer Vision for Games Workshop, San Francisco (USA), 2010.
[bibtex]

TL;DR. Conditioning particle filtering to run on surface of volumes ensure better tracking through a new occupancy resampling strategy.

Towards a Fair Evaluation of 3D Human Pose Estimation Algorithms
Cristian Canton, Josep R. Casas, Montse Pardàs, Enric Monte
Technical Report - Technical University of Catalonia, 2009.
[bibtex]

TL;DR. An approach to a fair estimation of human pose estimation through the Kolmogorov-Smirnov test.

Estimation of Head Pose
Michael Voit, Nicolas Gourier, Cristian Canton, Oswald Lanz,
Rainer Stiefelhagen, Roberto Brunelli
Book chapter - Computers in the Human Interaction Loop (Springer-Verlag), 2009.
[bibtex]

TL;DR. A survey of the technologies for head pose estimation in multimodal environments.

Activity classification
Kai Nickel, Montse Pardàs, Rainer Stiefelhagen, Cristian Canton,
José Luis Landabaso, Josep R. Casas
Book chapter - Computers in the Human Interaction Loop (Springer-Verlag), 2009.
[bibtex]

TL;DR. A survey on activity recognition technologies applied to SmartRooms.

Towards a Low Cost Multi-Camera Marker Based Human Motion Capture System
Cristian Canton, Josep R. Casas, Montse Pardàs
IEEE International Conference on Image Processing (ICIP), Cairo (Egypt), 2009.
[bibtex]  [poster]

TL;DR. Marker based motion tracking using annealed particle filtering.

Improving Detection of Acoustic Events Using Audiovisual Data and Feature Level Fusion
Taras Butko, Cristian Canton, Carlos Segura, Xavier Giró-i-Nieto, Climent Nadeu,
Javier Hernando, Josep R. Casas
Interspeech, Brighton (UK), 2009.
[bibtex]

TL;DR. Acoustic event detection using multiple cues and specific features using HMM.

Multimodal Acoustic Event Detection Towards Scene Understanding
Cristian Canton, Taras Butko, Carlos Segura, Xavier Giró-i-Nieto, Climent Nadeu,
Javier Hernando, Josep R. Casas
IEEE International Conference on Computer Vision and Pattern Recognition (CVPR). Human Communicative Behavior Analysis Workshop, Miami (USA), 2009.
[bibtex]  [poster]

TL;DR. Combination of motion cues, position and audio allows detection of events within a SmartRoom environment.

Voxel based Annealed Particle Filtering for Markerless 3D Articulated Motion Capture
Cristian Canton, Josep R. Casas, Montse Pardàs
IEEE Conference on 3DTV, Postdam (Germany), 2009.
[bibtex]

TL;DR. Markerless human body capture leveraging annealed particle filters prove to be effective and robust to occlusions.

Particle Filtering and Sparse Sampling for Multi-Person 3D Tracking
Cristian Canton, Rosella Sblendido, Josep R. Casas, Montse Pardàs
IEEE International Conference on Image Processing (ICIP), San Diego (USA), 2008.
[bibtex]  [poster]

TL;DR. Adding color to 3D voxel representations improve multi-person tracking.

Fusion of Audio and Video Modalities for Detection of Acoustic Events
Taras Butko, Andrey Temko, Climent Nadeu, Cristian Canton
Interspeech, Brisbane (Australia), 2008.
[bibtex]

TL;DR. Leveraging spatial information of people in a SmartRoom improves acoustic event detection.

Inclusion of Video Information for Detection of Acoustic Events using Fuzzy Integral
Taras Butko, Andrey Temko, Climent Nadeu, Cristian Canton
Workshop on Machine Learning and Multimodal Interaction (MLMI), Utrecht (The Netherlands), 2008.
[bibtex]

TL;DR. Multimodal data fusion towards acoustic events understanding in SmartRooms.

Exploiting Structural Hierarchy in Articulated Objects towards Robust Motion Capture
Cristian Canton, Josep R. Casas, Montse Pardàs
Conference on Articulated Motion and Deformable Objects (AMDO), Mallorca (Spain), 2008.
[bibtex]

TL;DR. New Montecarlo method applied to recover articulated structured (human body) exploiting hierarchies of complexity.

Multimodal Real-Time Focus of Attention Estimation in SmartRooms
Cristian Canton, Carlos Segura, Montse Pardàs, Josep R. Casas, Javier Hernando
IEEE International Conference on Computer Vision and Pattern Recognition (CVPR). Human Communicative Behavior Analysis Workshop, Anchorage (USA), 2008
[bibtex]  [poster]

TL;DR. Estimate the spatial focus of attention of a group of people leveraging their head orientation.

Audio-Driven Human Body Motion Analysis and Synthesis
Ferda Ofli, Cristian Canton, Joelle Tilmanne, Yasemir Demir, Elif Bozkurt,
Yucel Yemez, Engin Erzin, Ahmet Murat Tekalp
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Las Vegas (USA), 2008.
[bibtex]

TL;DR. Analysis and synthesis of motion patterns correlating audio and multiview information.

An Audio-Driven Dancing Avatar
Ferda Ofli, Cristian Canton, Joelle Tilmanne, Yasmin Demir,
Elif Bozkurt, Yucel Yemez, Engin Erzin, Ahmet Murat Tekalp
Journal - EURASIP Journal on Advances on Signal Processing, 2008.
[bibtex]

TL;DR. Analyze audio to make an avatar dance to the music.

Analysis and synthesis of multiview audio-visual dance figures
Ferda Ofli, Yasmin Demir, Cristian Canton, Joelle Tilmanne Tilmanne, Korai Balci,
Elif Bozkurt, I. Kizoglu, Yucel Yemez, Engin Erzin, Ahmet Murat Tekalp,
Lale Akarun, T.A. Erdem
IEEE Signal Processing, Communication and Applications Conference, Aydin (Turkey), 2008.
[bibtex]

TL;DR. Human dance patterns generation directly from music.

Flexible Test-bed for Unusual Behavior Detection
István Petrás, Levente Kovács, B. Uğur Töreyin, Csaba Beleznai, Zoltán Szlávik,
Uğur Güdükbay, Yiğithan Dedeoğlu, László Havasi, Enis Cetin, Montse Pardàs,
Tamás Szirányi, Cristian Canton
ACM International Conference on Image and Video Retrieval (CIVR), Amsterdam (The Nederlands), 2007.
[bibtex]

TL;DR. A system to combine multiple vision signals to detect unusual/outlier behaviors.

Head Orientation Estimation using Particle Filtering in Multiview Scenarios
Cristian Canton, Josep R. Casas, Montse Pardàs
Classification of Events, Activities and Relationships Evaluation and Workshop (CLEAR), Baltimore (USA), 2007.
[bibtex]

TL;DR. Define head templates and use them in a particle filtering context to estimate head position.

Multi-Person Tracking Strategies Based on Voxel Analysis
Cristian Canton, Jordi Salvador, Josep R. Casas
Classification of Events, Activities and Relationships Evaluation and Workshop (CLEAR), Baltimore (USA), 2007.
[bibtex]

TL;DR. Comparison of two tracking aproaches (heuristic vs particle filtering) on the CLEAR dataset.

Multi-Person 3D Tracking with Particle Filters on Voxels
Adolfo López, Cristian Canton, Josep R. Casas
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Honolulu (USA), 2007.
[bibtex]

TL;DR. Person tracking using particle filters directly on voxels, handling multiple objects and delivering high accuracy tracking.

Audiovisual Head Orientation Estimation with Particle Filters in Multisensor Scenarions
Cristian Canton, Carlos Segura, Josep R. Casas, Montse Pardàs, Javier Hernando
Journal - EURASIP Journal on Advances on Signal Processing, 2007.
[bibtex]

TL;DR. Monte Carlo techniques applied to audiovisual data fusion for head estimation and attention estimation purposes.

Multimodal Head Orientation Towards Attention Tracking in SmartRooms
Carlos Segura, Cristian Canton, Alberto Abad, Josep R. Casas, Javier Hernando
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Honolulu (USA), 2007.
[bibtex]

TL;DR. Head pose estimation using both multicamera and microphone arrays signals increase robustness and accuracy of estimation.

Human Model and Motion Based 3D Action Recognition in Multiple View Scenarios
Cristian Canton, Josep R. Casas, Montse Pardàs
European Signal Processing Conference (EUSIPCO), Florence (Italy), 2006.
[bibtex]   [slides]

TL;DR. Analyze 3D volumes using a human body model to recognize actions in multicamera scenarios.

3D Human Action Recognition in Multiple View Scenarios
Cristian Canton, Josep R. Casas, Montse Pardàs, Mehmet Emre Sargin, Ahmet Murat Tekalp,
Jornades de Recerca en Automàtica, Visió i Robòtica (AVR), Barcelona (Spain), 2006.
[bibtex]

TL;DR. Use multiple cameras to create voxels and recognize actions directly from the 3D.

Head Pose Detection Based on Fusion of Multiple Viewpoint Information
Cristian Canton, Josep R. Casas, Montse Pardàs
Classification of Events, Activities and Relationships Evaluation and Workshop (CLEAR), Southampton (UK), 2006.
[bibtex]

TL;DR. Head position estimation using multiview face appearance and its evaluation against the CLEAR dataset.

UPC Audio, Video and Multimodal Person Tracking Systems in the CLEAR Evaluation Campaign
Alberto Abad, Cristian Canton, Carlos Segura, José Luis Landabaso, Dusan Macho,
Josep R. Casas, Javier Hernando, Montse Pardàs, Climent Nadeu
Classification of Events, Activities and Relationships Evaluation and Workshop (CLEAR), Southampton (UK), 2006.
[bibtex]

TL;DR. Multiple modalities always improve tracking, specially of humans within a SmartRoom.

Fusion of Multiple Viewpoint Information Towards 3D Face Robust Orientation Detection
Cristian Canton, Josep R. Casas, Montse Pardàs
IEEE International Conference on Image Processing (ICIP), Genoa (Italy), 2005.
[bibtex]

TL;DR. Head orientation estimation by 3D backpropagation of detected skin patches on multiple views.

Projective Kalman Filter: Multiocular Tracking of 3D Locations Towards Scene Understanding
Cristian Canton, Josep R. Casas, Ahmet Murat Tekalp, Montse Pardàs
Workshop on Multimodal Interaction and Related Machine Learning Algorithms (MLMI), Edinburgh (UK), 2005.
[bibtex]

TL;DR. Kalman filtering of 3D points directly from 2D projections; robust to occlusions by analyzing outliers. Simple yet effective!

Smart Spaces: Recerca en Interfícies Multimodals per afrontar els reptes de les tecnologies del futur (in Catalan)
Josep R. Casas, Cristian Canton
Telecomunications and Electronics Forum Magazine, Technical University of Catalonia, 2005.
[bibtex]

TL;DR. Multimodal understanding in SmartRooms will enable HCI scenarios to make meetings more effective.

Towards a Bayesian Approach to Robust Finding Correspondences in Multiple View Geometry Environments
Cristian Canton, Josep R. Casas, Montse Pardàs
International Conference on Computer Science (ICCS). Workshop on Computer Graphics and Geometric Modelling, Atlanta (USA), 2005.
[bibtex]

TL;DR. Tracking multiple 3D points in a multiview scenario using a greedy Bayesian approach. Useful for low compute scenarios and SmartRooms.

Evaluation of the Efficiency of Robust Anisotropic Diffusion Schemes
Bogdan Smolka, Cristian Canton, Marek Szczepanski, Konrad Wojciechowski
International Conference on Computer Vision and Graphics (ICCVG), Zakopane (Poland), 2002.
[bibtex]

TL;DR. Conductance functions used in anisotropic diffusion schemes may have relevant impact on the final result.

Fingerprint representation and matching using redundant expansions
Cristian Canton
Technical Report, Technical University of Catalonia, 2003.
[bibtex]

TL;DR. Use overcomplete representations to represent fingerprints and match them against others.

Efficient Algorithm for Impulsive Noise Reduction
Bogdan Smolka, Cristian Canton
IFAC Workshop on Programable Devices and Systems, Gliwice (Poland), 2001.
[bibtex]

TL;DR. An old shcool image filtering technique based on the maximization of the similarities between pixels in the filtering window.

On the Efficiency of Anisotropic Diffusion Filtering Schemes
Bogdan Smolka, Cristian Canton, Marko Marcevski, Zeina Torbey
IFAC Workshop on Programable Devices and Systems, Gliwice (Poland), 2001.
[bibtex]

TL;DR. When running anisotropic diffusion filters on images, not all conduction functions were born equal.

Theses
Human Motion Capture with Scalable Body Models , PhD Thesis
Cristian Canton
Technical University of Catalonia, Barcelona (Spain), 2009.
[bibtex]  [PDF (High Resolution)]

TL;DR. Human motion capture using hierarchical models and Monte Carlo optimization techniques. No fancy ML involved, just math and pixels.

Texture and Edge Adaptive Weak Matching Pursuit , MS Thesis
Cristian Canton
École Polytechnique Fédérale de Lausanne (EPFL), Lausanne (Switzerland), 2003.
[bibtex]

TL;DR. Overcomplete representations allow high compression of images/videos. Exploiting texture and edge priors into the process yield to dramatic improvements.

Service
Area Chair, ICCV 2021
General chair, Workshop on Media Forensics @ CVPR 2021
Program Chair, WACV 2021 Co-chair, DeepVision Workshop @ CVPR 2020
General chair, Workshop on Media Forensics @ CVPR 2020
Co-chair, DeepVision Workshop @ CVPR 2019
General chair, Workshop on Media Forensics @ CVPR 2019
General chair, Workshop on Objectionable Content and Misinformation @ ECCV 2018
Musicology research

As unusual as it may seem, aside of science/AI/CV I am also a pianist and musicologist specialized in the Catalan and Spanish composers that immigrated to America in the XIXth and early XXth centuries. Yes, everyone has a weird passion.

Jaime Nunó: más allá del Himno Nacional Mexicano (in Spanish)
Cristian Canton, Raquel Tovar
Book, USA and Mexico, 2017.
[bibtex]

TL;DR. Second edition (revised and extended) presenting the biography of Catalan composer Jaime Nunó, composer of Mexico's national anthem.

Jaime Nunó: su legado musical (in Spanish)
Cristian Canton
Book - Mozaic Editions and Fondo Nacional para la Cultura y las Artes (FONCA), UK and Mexico, 2012.
[bibtex]

TL;DR. First critical edition of the complete works of Jaime Nunó (1824-1906) from his manuscripts and first editions.

Luis G. Jordà: un músico catalán en el México Porfiriano (in Spanish)
Cristian Canton
Book - Mozaic Editions and Fondo Nacional para la Cultura y las Artes (FONCA), UK and Mexico, 2011.
[bibtex]

TL;DR. First biography of Catalan composer Luis G. Jorda (1869-1951) and his trajectory in America and Catalonia during late XIX and early XX centuries.

Jaime Nunó. Un sanjuanense en América (in Spanish)
Cristian Canton, Raquel Tovar
Book - Casa Amèrica-Catalunya, Consulado General de México en Barcelona, Ajuntament de Sant Joan de les Abadesses, Spain, 2010.
[bibtex]

TL;DR. First biography (in Spanish) of the composer of Mexico's national anthem: the Catalan musician Jaime Nunó (1824-1906).

Jaume Nunó. Un santjoaní a Amèrica (in Catalan)
Cristian Canton, Raquel Tovar
Book - Casa Amèrica-Catalunya, Consulado General de México en Barcelona, Ajuntament de Sant Joan de les Abadesses, Spain, 2010.
[bibtex]

TL;DR. First biography (in Catalan) of the composer of Mexico's national anthem: the Catalan musician Jaime Nunó (1824-1906).

Vida i Obra de Luis G. Jordà (in Catalan)
Cristian Canton
Book - Ajuntament de Les Masies de Roda, Spain, 2010.
[bibtex]

TL;DR. First biography of Catalan composer Luis G. Jorda (1869-1951) and his trajectory in America and Catalonia during late XIX and early XX centuries.

Música en los tiempos del Centenario de la Independencia de México (1910) (in Spanish)
Cristian Canton et al.
Book - Sextil Editores, Mexico, 2009.
[bibtex]

TL;DR. An in-depth review of the music in Mexico during the hundreth anniversary of its independence.


Like the style? Shamelessly borrowed from Jon Barron's site.