Silvia Zuffi

Phd in Computer Science, Brown University, May 2015 PhD Thesis: Shape Models of the Human Body for Distributed Inference. Advisor: Prof. Michael J. Black Master degree in Computer Science, Brown University, May 2010 Master degree in Electronic Engineering at the Faculty of Electronic Engineering of the University of Bologna, March 1995. Thesis (in Italian): Applicazione di un benchmark di elaborazione delle immagini ad una macchina parallela floating point e analisi delle prestazioni (Application of the DARPA Image Understanding Benchmark to a parallel computer with SIMD floating point architecture and performance analysis). Supervisor: Prof. Tullio Salmon Cinotti, 100/100 magna cum laude.

Research Activity

My current research interest is in animal and human shape and pose estimation from images and videos. My main contribution to the field is the exploration of realistic 2D and 3D models of body shape for estimating pose and articulated motion in unconstrained images, videos, and from 3D data. In the past I have worked on color imaging and reproduction, in particular on the readability of colored text on colored background and on multispectral print.


New paper at CVPR 2023 "BITE: Beyond Priors for Improved Three-D Dog Pose Estimation" Project page

BARC: Breed-Augmented Regression Using Classification for 3D Dog Reconstruction from Images published on IJCV PDF

CV4animals workshop at CVPR 2023 Website

New paper at CVPR 2022 on modeling and inferring the human body skeleton: "OSSO: Obtaining Skeletal Shape from Outside" Project page

New paper at CVPR 2022 on estimating 3D dogs from images: "BARC: Learning to Regress 3D Dog Shape from Images by Exploiting Breed Information" Project page

Perspectives in machine learning for wildlife conservation Nature Communications

New work on sound separation applied to fish vocalizations

New paper on Biodiversity Monitoring accepted at Nature Communications, preprint available

I will speak at the CV4Ecology summer school at Caltech

Recordings of the Invited Talks at CV4Animals are available

New "work in progress" publication: Ci Li, Nima Ghorbani, Sofia Broome, Maheen Rashid, Michael J. Black, Elin Hernlund, Hedvig Kjellström, Silvia Zuffi, "hSMAL: Detailed Horse Shape and Pose Reconstruction for Motion Pattern Recognition", CVPR 2021, CV4Animals Workshop (oral),

I will be Program Chair at 3DV2021

List of publications including those not associated with CNR projects:


Blog post on animal models

Ongoing Projects



2023, Journal article
BARC: Breed-Augmented regression using classification for 3D dog reconstruction from images
N. Rueegg, S. Zuffi, K. Schindler, and M.J. Black
The goal of this work is to reconstruct 3D dogs from monocular images. We take a model-based approach, where we estimate the shape and pose parameters of a 3D articulated shape model for dogs. We cons...

CNR@People | DOI: 10.1007/s11263-023-01780-3
2022, Journal article
Perspectives in machine learning for wildlife conservation
D. Tuia, B. Kellenberger, S. Beery, B.R. Costelloe, S. Zuffi, B. Risse, A. Mathis, M.W. Mathis, F. van Langevelde, T. Burghardt, R. Kays, H. Klinck, M. Wikelski, I.D. Couzin, G. van Horn, M. Crofoot, C.V. Stewart, and T. Berger-Wolf
Inexpensive and accessible sensors are accelerating data acquisition in animal ecology. These technologies hold great potential for large-scale ecological understanding, but are limited by current pro...

CNR@People | DOI: 10.1038/s41467-022-27980-y
2016, Journal article
Body talk: Crowdshaping realistic 3D avatars with words
S. Streuber, M.A. Quiros-Ramirez, M.Q. Hill, C.A. Hahn, S. Zuffi, A. O'Toole, M.J. Black
Realistic, metrically accurate, 3D human avatars are useful for games, shopping, virtual reality, and health applications. Such avatars are not in wide use because solutions for creating them from hig...

CNR@People | DOI: 10.1145/2897824.2925981
2008, Journal article
From color sensor space to feasible reflectance spectra
Zuffi S; Santini S; Schettini R
The interaction of light and object surfaces generates color signals in the visible band that are responsible for digital ac- quisition system outputs. Inverting this mapping from the sensor space bac...

CNR@People | DOI: 10.1109/TSP.2007.907838
2008, Journal article
Uno studio sulla corrispondenza tra esperimenti controllati e non controllati
S. Zuffi, C. Brambilla, R. Eschbach, A. Rizzi,

2007, Journal article
A Computational Strategy Exploiting Genetic Algorithms to Recover Color Surface Reflectance Functions
R. Schettini, S. Zuffi
Information about the spectral reflectance of a color surface is useful in many applications. Assuming that reflectance functions can be adequately approxi- mated by a linear combination of a small nu...

CNR@People | DOI: 10.1007/s00521-006-0049-7
2007, Journal article
Turismo Sistema Produttivo Aperto
V. Esposti; I. Gagliardi; E. Oliveri; M. Sacco; B. Zonta; S. Zuffi
Il turismo è un settore trasversale che da un lato potrebbe diventare un potente motore del ciclo conservazione - valorizzazione - fruizione del patrimonio culturale/ambientale, rappresentando ...

2006, Journal article
Accounting from Inks Interaction in the Yule Nielsen Spectral Neugebauer Model
S. Zuffi, S. Santini, R. Schettini,
Multispectral printer characterization requires an effec-tive model to map the inputs to the printer (i.e., the digital counts ofthe inks) into reflectance spectra and vice versa. Most of the meth-ods...

CNR@People | DOI: 10.2352/J.ImagingSci.Technol.(2006)50:1(35)
2005, Journal article
Spectral-based printer modeling and characterization
Zuffi S., Schettini R., Mauri G.
Spectral characterization involves building a model that relates the device dependent representation to the reflectance func- tion of the printed color, usually represented with a high number of refle...

CNR@People | DOI: 10.1117/1.1900137
2011, Essay or book chapter
Applications of spectral imaging and reproduction to cultural heritage
S. Bianco, A. Colombo, F. Gasparini, R. Schettini, S. Zuffi

2022, Conference proceedings
BARC: Learning to regress 3D dog shape from images by exploiting breed information
N. Rueegg, s. Zuffi, K. Schindler, and M.J. Black
Our goal is to recover the 3D shape and pose of dogs from a single image. This is a challenging task because dogs exhibit a wide range of shapes and appearances, and are highly articulated. Recent wor...

CNR@People | DOI: 10.1109/CVPR52688.2022.00385
2022, Conference proceedings
OSSO: Obtaining skeletal shape from outside
M. Keller, S. Zuffi, M.J. Black, and S. Pujades
We address the problem of inferring the anatomic skeleton of a person, in an arbitrary pose, from the 3D surface of the body; i.e. we predict the inside (bones) from the outside (skin). This has many ...

CNR@People | DOI: 10.1109/CVPR52688.2022.01984
2019, Conference proceedings
Three-D Safari: Learning to estimate zebra pose, shape, and texture from images "In the Wild"
S. Zuffi, A. Kanazawa, T. Berger-Wolf, and M.J. Black
We present the first method to perform automatic 3Dpose, shape and texture capture of animals from imagesacquired in-the-wild. In particular, we focus on the problemof capturing 3D information about G...

CNR@People | DOI: 10.1109/ICCV.2019.00546
2018, Conference proceedings
Lions and tigers and bears: Capturing non-rigid, 3D, articulated shape from images
S. Zuffi, A. Kanazawa, and M.J. Black
Animals are widespread in nature and the analysis of their shape and motion is important in many fields and industries.Modeling 3D animal shape, however, is difficult because the 3D scanning methods u...

CNR@People | DOI: 10.1109/CVPR.2018.00416
2017, Conference proceedings
3D menagerie: Modeling the 3D shape and pose of animals
S. Zuffi, A. Kanazawa, D. Jacobs, and M.J. Black
There has been significant work on learning realistic, articulated, 3D models of the human body. In contrast, there are few such models of animals, despite many applications. The main challenge is tha...

CNR@People | DOI: 10.1109/CVPR.2017.586
2015, Conference proceedings
The Stitched Puppet: A Graphical Model of 3D Human Shape and Pose
S. Zuffi and M.J. Black
We propose a new 3D model of the human body that isboth realistic and part-based. The body is represented by agraphical model in which nodes of the graph correspond tobody parts that can independently...

CNR@People | DOI: 10.1109/CVPR.2015.7298976
2013, Conference proceedings
Estimating Human Pose with Flowing Puppets
S. Zuffi, J. Romero, C. Schmid, M. J. Black

CNR@People | DOI: 10.1109/ICCV.2013.411
2013, Conference proceedings
Towards understanding action recognition
H. Jhuang, J. Gall, S. Zuffi, C. Schmid, and M. J. Black

2012, Conference proceedings
From Pictorial Structures to Deformable Structures
Silvia Zuffi, Oren Freifeld, Michael J. Black
Pictorial Structures (PS) define a probabilistic model of 2D articulated objects in images.Typical PS models assume an object can be represented by a set of rigid parts connected withpairwise constrai...

2010, Conference proceedings
Contour people: a paremeterized model of 2D articulated human shape
O. Freifeld, A. Weiss, S. Zuffi, M. J. Black
We define a new "contour person" model of the human body that has the expressive power of a detailed 3D model and the computational benefits of a simple 2D part-based model. The contour pers...

CNR@People | DOI: 10.1109/CVPR.2010.5540154
2009, Conference proceedings
A study on the equivalence of controlled and uncontrolled visual experiments
Zuffi S., Brambilla C., Eschbach R., Rizzi A.

2008, Conference proceedings
Controlled and Uncontrolled Viewing Conditions in the Evaluation of Prints
S. Zuffi, C. Brambilla, R. Eschbach, A. Rizzi

2008, Conference proceedings
Valutazione di immagini in condizioni controllate e non controllate di osservazione
Zuffi S., Brambilla C., Eschbach R., Rizzi A.

2008, Conference proceedings
Visual Experiments on the Web: Design of a Web-based Visual Experiment Management System
S. Zuffi, E. Beltrame, P.L. Scala

2007, Conference proceedings
Comparison of Techniques for Mitigating the Effects of Illumination Variations on the Appearance of Human Targets
C. Madden, M. Piccardi, S. Zuffi
Several techniques have been proposed to date to build colour invariants between camera views with varying illumination con- ditions. In this paper, we propose to improve colour invariance by using da...

CNR@People | DOI: 10.1007/978-3-540-76856-2_12
2007, Conference proceedings
Human computer interaction:legibility and contrast
Zuffi S.; Brambilla C.; Beretta G.; Scala P.
In human computer interaction readability of textual information is one of the foremost requirements for displays designed to provide visual information. In this context it is of main interest to iden...

CNR@People | DOI: 10.1109/ICIAP.2007.4362786
2007, Conference proceedings
Valutazione automatica di interventi di restauro del colore in opere cinematografiche
S. Zuffi, R. Cereda, A. Rizzi

2007, Conference proceedings
Web-based vs. controlled environment psychophysics experiments
Zuffi S., Brambilla C., Beretta G., Scala P.

2006, Conference proceedings
A color interface for audio clustering visualization
Zuffi S., Gagliardi I
The availability of large audio collections calls for ways to efficiently access and explore them by providing an effective overview of their contents at the interface level. In this paper we present ...

CNR@People | DOI: 10.1117/12.649210
2006, Conference proceedings
A Color Selection Tool for the Readability of Textual Information on Web pages
Zuffi S., G. Beretta, C. Brambilla

2006, Conference proceedings
Promoting Cultural Tourism across Mediterranean Countries through ICT technologies: The Daedalus Project
De Massis A, Della Ventura A., T. Karathanasis, G. Tosi, S. Zuffi,


Go BackImati logo