17.08.2013 Views

Mr Kiran Varanasi / Dr Fabio Cuzzolin Applicant Career Summary

Mr Kiran Varanasi / Dr Fabio Cuzzolin Applicant Career Summary

Mr Kiran Varanasi / Dr Fabio Cuzzolin Applicant Career Summary

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong><br />

Title:<br />

First Name:<br />

Surname:<br />

Other Names:<br />

Honours:<br />

Address:<br />

Town:<br />

Postcode:<br />

Country:<br />

Nationality:<br />

Email Address:<br />

Telephone (work):<br />

Fax:<br />

Abstract:<br />

<strong>Mr</strong><br />

<strong>Kiran</strong><br />

<strong>Varanasi</strong><br />

655 Avenue de l'Europe<br />

Montbonnot<br />

St Ismier<br />

38334<br />

France<br />

Nationality<br />

Indian<br />

<strong>Applicant</strong> <strong>Career</strong> <strong>Summary</strong><br />

INRIA Grenoble Rhone Alpes<br />

vakibs@gmail.com<br />

+33-616574493<br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

Newton International Fellowships - 2010<br />

As cameras and depth sensors get cheaper, they become deployed ubiquitously<br />

and capture images and video that are large beyond any human consumption.<br />

These are meant to be processed by computer algorithms, that analyse and<br />

summarize the information for human beings. Several challenges remain for<br />

developing such algorithms, especially in analysing motion and understanding<br />

scene dynamics.<br />

When it is known beforehand the kind of objects that are being observed, a<br />

particular template model can be chosen and fit to the observed image (e.g,<br />

Project Natal of Microsoft). However, such an approach cannot handle unknown<br />

scenes, when multiple actors are present in an outdoor environment.<br />

Unsupervised learning is necessary in that case. We propose to use "manifold<br />

learning", a statistical framework, to achieve this. This has to be extended to<br />

partial recognition, when objects in the scene are not segmented from<br />

background. Applications include cinematic motion capture and automatic<br />

surveillance.<br />

Page 1 of 8


<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong> Newton International Fellowships - 2010<br />

Statement of<br />

qualifications and<br />

career:<br />

Field of Specialisation:<br />

Publications:<br />

Subject:<br />

Present Research:<br />

Present Position:<br />

Present Employer:<br />

Research Assistant, Center for Visual Information<br />

Technology, IIIT Hyderabad, India<br />

Principal Mentor (Instructor for computer graphics<br />

course), MSIT Program, India<br />

Research Associate, ISRI, Carnegie Mellon University,<br />

USA<br />

INRIA Grenoble Rhone Alpes<br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

Qualification Date<br />

Ph.D Fellow, INRIA Rhône Alpes, Grenoble, France 26/11/2006 - 25/05/2010<br />

01/04/2004 - 31/10/2006<br />

01/08/2004 - 31/03/2005<br />

01/06/2003 - 31/12/2003<br />

Computer Vision, Spatio-temporal modeling of dynamic scenes, 3D features for<br />

matching and recognition, Tracking from images<br />

"Temporal surface tracking via mesh evolution"<br />

<strong>Kiran</strong> <strong>Varanasi</strong>, Andrei Zaharescu, Edmond Boyer, Radu Horaud<br />

European Conference on Computer Vision (ECCV), Marseille, October 2008<br />

(Oral Presentation : 5% acceptance rate)<br />

"Surface feature detection and description with application to mesh matching"<br />

Andrei Zaharescu, Edmond Boyer, <strong>Kiran</strong> <strong>Varanasi</strong>, Radu Horaud<br />

International Conference on Computer Vision and Pattern Recognition (CVPR),<br />

Miami, June 2009<br />

"A document space model for automated text classification based on frequency<br />

distribution across categories"<br />

<strong>Kiran</strong> <strong>Varanasi</strong>, Chaitanya Kamisetty, Sushma Bendre, Rajeev Sangal, Akshar<br />

Bharati<br />

International Conference on Natural Language Processing (ICON), Mumbai,<br />

December 2002<br />

NIF Group 05: Information communication technology (ICT) / Computer Vision -<br />

ICT<br />

Through synchronized cameras in an indoor setting, a multi-view video is captured<br />

of different actors performing diverse tasks and interacting with each other. 3D<br />

mesh-models of the scene are built independently, by extracting silhouettes at<br />

each frame. These meshes are not topologically or geometrically coherent, due to<br />

artefacts in silhouette extractions and by occlusions. I have worked on obtaining a<br />

coherent spatio-temporal model of the observed scene, which connects these<br />

meshes.<br />

I worked on dense 3D motion estimation, which is a particularly hard problem<br />

when no assumptions can be made of the scene, as that it has a single topology<br />

over time. I demonstrated results on complex sequences, e.g, a dancer wearing a<br />

loose fitting robe with a feather boa belt, a juggler juggling with clubs, two children<br />

playing ball with each other etc. No additional information was input, apart from<br />

that captured by the cameras. I achieved this through a mesh evolution framework<br />

which accounts for topological changes.<br />

Later, I worked on computing 3D features directly on meshes, which can be used<br />

for initializing global motion estimation.<br />

Finally, I worked on a temporally coherent segmentation scheme to automatically<br />

extract body-parts that move rigidly over time. I demonstrated results on similar<br />

complex sequences involving multiple actors and background clutter (recent work<br />

submitted for publication).<br />

Thus, spatio-temporal modelling of arbitrary scenes is achieved at various scales.<br />

Doctoral Student<br />

Page 2 of 8


<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong> Newton International Fellowships - 2010<br />

Present Department:<br />

Present Position Start<br />

Date:<br />

Present Position End<br />

Date:<br />

PhD Awarded Date:<br />

PhD Institution:<br />

PhD Country:<br />

Previous Support<br />

Description:<br />

Where did you hear of<br />

this scheme?:<br />

26/11/2006<br />

25/05/2010<br />

Co-<strong>Applicant</strong> <strong>Career</strong> <strong>Summary</strong><br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

Perception Group (Computer Science Department)<br />

30/06/2010<br />

INRIA Grenoble Rhône Alpes and Laboratoire Jean Kuntzmann, Université<br />

Grenoble<br />

France<br />

Employed by INRIA Grenoble. Supported by the INTERACT project of<br />

PERCEPTION group (European Grant)<br />

Other<br />

Co-<strong>Applicant</strong> Personal Details<br />

Title:<br />

First Name:<br />

Surname:<br />

Other Names:<br />

Honours:<br />

Address Line 1:<br />

Address Line 2:<br />

Address Line 3:<br />

Address Line 4:<br />

Address Line 5:<br />

Town:<br />

Postcode:<br />

Country:<br />

Nationality:<br />

Email Address:<br />

Telephone (work):<br />

Fax:<br />

<strong>Dr</strong><br />

<strong>Fabio</strong><br />

<strong>Cuzzolin</strong><br />

Ph.D.<br />

Department of Computing<br />

Oxford Brookes University<br />

Wheatley campus<br />

Wheatley<br />

OX33 1HX<br />

United Kingdom<br />

Nationality<br />

Italian<br />

fabio.cuzzolin@brookes.ac.uk<br />

01865 484526<br />

01865 484545<br />

Page 3 of 8


<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong> Newton International Fellowships - 2010<br />

Statement of<br />

qualifications and<br />

career:<br />

Field of Specialisation:<br />

Publications:<br />

Subject:<br />

Present Research:<br />

Present Position:<br />

Present Employer:<br />

Present Department:<br />

Marie Curie Fellow, INRIA Rhone-Alpes, Grenoble,<br />

France<br />

Post-doctoral researcher, University of California at<br />

Los Angeles<br />

Fixed-term assistant professor, Politecnico di Milano,<br />

Italy<br />

Lecturer - Early <strong>Career</strong> Fellow<br />

Oxford Brookes University<br />

Department of Computing<br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

Qualification Date<br />

Lecturer, Oxford Brookes University 01/09/2008 - to date<br />

04/09/2006 - 03/09/2008<br />

01/10/2004-09/04/2006<br />

01/01/2003-31/12/2004<br />

Post-doctoral researcher, University of Padua, Italy 01/06/2001-31/05/2003<br />

<strong>Dr</strong> <strong>Cuzzolin</strong>’s research interests include machine and manifold learning, computer<br />

vision and human motion analysis, the theory of belief functions and imprecise<br />

probabilities.<br />

<strong>Fabio</strong> <strong>Cuzzolin</strong>, Multilinear modeling for robust identity recognition from gait, in<br />

“Behavioral Biometrics for Human Identification: Intelligent Applications", Liang<br />

Wang and Xin Geng (Eds.), IGI Publishing, 2009<br />

<strong>Fabio</strong> <strong>Cuzzolin</strong>, A geometric approach to the theory of evidence, IEEE<br />

Transactions on Systems, Man, and Cybernetics part C, 38(4), pages 522-534,<br />

July 2008<br />

<strong>Fabio</strong> <strong>Cuzzolin</strong>, Diana Mateus, David Knossow, Edmond Boyer, and Radu<br />

Horaud, Coherent Laplacian protrusion segmentation, Proceedings of CVPR'08,<br />

Anchorage, Alaska;<br />

Diana Mateus, Radu Horaud, David Knossow, <strong>Fabio</strong> <strong>Cuzzolin</strong>, and Edmond<br />

Boyer, Articulated Shape Matching Using Laplacian Eigenfunctions and<br />

Unsupervised Point Registration, Proceedings of CVPR'08, Anchorage, Alaska;<br />

<strong>Fabio</strong> <strong>Cuzzolin</strong>, Using Bilinear Models for View-invariant Action and Identity<br />

Recognition, Proceedings of<br />

CVPR'06, pp. 1701-1708, New York, June 18-22 2006<br />

NIF Group 05: Information communication technology (ICT) / Computer Vision -<br />

ICT<br />

<strong>Dr</strong> <strong>Cuzzolin</strong>’s research interests include machine learning, computer vision and<br />

imprecise probabilities.<br />

He is first or single author of some 50 publications (including 9 journals + 6 under<br />

review), some of which received awards. He collaborates with several journals in<br />

both computer vision and probabilities, and served in the program committee of<br />

some 15 international conferences.<br />

<strong>Dr</strong> <strong>Cuzzolin</strong> is a prominent expert in the field of random sets. He formulated a<br />

geometric approach to uncertainty in which probabilities, possibilities and belief<br />

functions can all be represented as points of a Cartesian space and there<br />

analyzed. He studied how to approximate random sets with probabilities, and<br />

proposed novel formulations of the theory of belief functions.<br />

Within computer vision, his work focused on human motion analysis and action<br />

recognition. He proposed the use of multilinear models for identity recognition<br />

from gait, and explored spectral motion capture techniques for unsupervised 3D<br />

segmentation and matching.<br />

<strong>Dr</strong> <strong>Cuzzolin</strong> is finalizing collaborations with IDSIA, Switzerland for a STREP on<br />

imprecise Markov chains for gesture recognition, and with INRIA, Pompeu Fabra<br />

and Technion on a Future and Emerging Technology (FET) EU proposal on large<br />

scale manifold learning. He is discussing a collaborative project on uncertainty<br />

theory at UK level with U. Bristol and Durham’s Dept of Statistics. He is also<br />

exploring the opportunity of a European Network of Excellence in the same fiel<br />

Page 4 of 8


<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong> Newton International Fellowships - 2010<br />

Present Position Start<br />

Date:<br />

Present Position End<br />

Date:<br />

PhD Awarded Date:<br />

Proposal<br />

Current Funding<br />

Description:<br />

Previous Support<br />

Description:<br />

Subject:<br />

Project Title:<br />

Start Date:<br />

End Date:<br />

Research proposal:<br />

01/09/2008<br />

31/08/2013<br />

19/02/2001<br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

<strong>Dr</strong> <strong>Cuzzolin</strong> is currently applying for his EPSRC First Grant, he is finalizing a FET<br />

(Future Emerging Technology) EU FP7 project with as partners INRIA Rhone-<br />

Alpes, U. Pompeu Fabra (Barcelona), and Technion (Haifa). He is also preparing<br />

a Leverhulme research proposal, and will submit a European Research Council<br />

Starting Grant before October 2010. No funding though is currently available.<br />

<strong>Dr</strong> <strong>Cuzzolin</strong> has received Oxford Brookes' central university funding in 2008<br />

(some 3600 pounds overall).<br />

NIF Group 05: Information communication technology (ICT) / Computer Vision -<br />

ICT<br />

Manifold learning for motion analysis and recognition<br />

01/01/2011<br />

31/12/2012<br />

In the course of this project, we will explore the problem of recognizing objects<br />

and their motion in an unknown environment filled with background clutter. One<br />

good example is the recognition of human activities in an outdoor environment,<br />

from a single image or from a set of images, obtained from cameras or depth<br />

sensors. In this case, we aim to recognize the human beings in the scene, their<br />

body postures and their motion over time. Our problem is made complex by the<br />

facts that the people need not be visible entirely to the camera, that they might<br />

wear loose clothes which cause further occlusions, and that they might interact<br />

with each other using various unknown objects.<br />

In contrast, if it is known that a person faces straight to the sensor, or at least, that<br />

he/she is visible entirely to the sensor's field of view, a known articulated model of<br />

human beings can be fit to the observed image and the pose of the body be<br />

computed. Certain commercial applications are in the course of development (e.g,<br />

Project Natal of Microsoft corporation) which use cameras and depth sensors in<br />

this way, to compute body poses and provide novel modes of interaction for users<br />

to play computer games. The variations in human body shapes and their poses<br />

are learnt into a statistical model, using a vast database of actors captured in a<br />

pre-defined set of poses. Apart from commercial ventures, there have also been<br />

various research publications that studied this problem in the recent past. Diverse<br />

solutions have been proposed to define and represent such statistical model, and<br />

to compute the closest fit for an observation.<br />

One such solution was proposed by <strong>Dr</strong>. <strong>Cuzzolin</strong> et al, through manifold learning.<br />

The observation is transformed into a new space using adaptive spectral<br />

embeddings that are isometrically invariant (i.e, that are invariant to articulated<br />

motions). Articulated motion can be estimated as a simple global rotation in this<br />

embedding space. The power of this solution is its generality, as it can be used for<br />

motion estimation of an unknown object, not necessarily a human being.<br />

However, its weakness is that the object needs to be segmented properly from the<br />

background, which is difficult to achieve in a natural setting, or when multiple<br />

actors are present.<br />

Page 5 of 8


<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong> Newton International Fellowships - 2010<br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

During my Ph.D, I have worked on the problem of motion estimation without<br />

segmenting an object from background clutter. I demonstrated results on scenes<br />

involving multiple actors interacting with each other. To achieve this, I used a<br />

mesh evolution framework that handles topological changes. I proposed methods<br />

for computing local 3D features on objects, and algorithms for matching shapes<br />

that rely on such local features. Such methods can handle partial shape matching,<br />

and work even when the object is not entirely visible to the sensor.<br />

In the current project, we aim at combining the strengths of the two approaches<br />

developed by applicant and co-applicant and fit them into a statistical learning<br />

framework. The first challenge that we face is the huge dimensionality of the input<br />

data. Even though adaptive spectral embeddings translate this input into a new<br />

space of much lower dimensionality, they have to be first learnt by training. When<br />

the dimensionality of the input data is prohibitively large (as is in our case), the<br />

learning takes too much time and needs too much memory space. We would like<br />

to attack this problem by using the idea of "compressed sensing" that has recently<br />

become popular in the signal processing community. We shall code the input data<br />

using random projections that are linear, sparsity oriented and metric-preserving.<br />

This vastly reduces the dimensionality of the feature space and makes adaptive<br />

embeddings tractable.<br />

Adaptive spectral embeddings can handle global matching, but not partial<br />

matching. Our second challenge is to fit these embeddings into a partial matching<br />

framework. Similar to my Ph.D work, we would like to derive feature descriptors<br />

on these embeddings that are local in scope, and which can be used for matching<br />

shapes amidst background clutter. Ideally, we should be able to learn body-parts<br />

of the shapes, with their distinctive shape signatures.<br />

The third and final challenge that we will have to address is that of unsupervised<br />

learning over a large database of example shapes and poses. This learning will<br />

further improve the descriptiveness of our features. For example, statistical<br />

learning can restrict the lower arm of a human body to have a joint angle not more<br />

than 180 degrees with the upper arm. We seek to learn the shape and motion<br />

statistics of arbitrary shapes, not just human beings. For example, we can learn<br />

the statistics of men riding bicycles, or of children playing with their pets.<br />

The potential applications of the proposed work are numerous. A straightforward<br />

application, for instance, is kinematic motion capture in outdoor environments,<br />

crucial in the entertainment industry, such as in film production and gaming. A<br />

second natural application is automatic visual surveillance, especially in crowded<br />

and cluttered scenes where the suspect is on the run. An additional example<br />

concerns the assistance of elderly people, through monitoring for signs of ageing<br />

and weakness in their limb movements. A fourth application could be in sports<br />

analysis, where the movements of a sports-person are analysed for their<br />

correctness. Due to the general-purpose nature of our approach, it shall be able<br />

for deployment amidst multiple people performing diverse tasks, and in outdoor<br />

environments. This opens doors for several new applications, to a degree that is<br />

not possible with today's technology.<br />

Apart from practical applications, there shall also be theoretical breakthroughs that<br />

would be of interest, for various other fields that fit in the broad scope of pervasive<br />

computing. Bringing together the two fields of manifold learning and compressed<br />

sensing fits the broader research goal of <strong>Dr</strong>. <strong>Cuzzolin</strong>, as outlined by his<br />

collaboration with other European centers of research. The task of analysing<br />

motion (and human activities in particular) fits well with the long-term goals of the<br />

Vision Group of Oxford Brookes University, which has produced several worldclass<br />

publications in this field. I would like to benefit from the strong collaborations<br />

that tie the group in Oxford Brookes to other centres of excellence in the world,<br />

particularly to the Visual Geometry group in the University of Oxford and to the<br />

Page 6 of 8


<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong> Newton International Fellowships - 2010<br />

Accompanying<br />

dependants:<br />

Previous contact:<br />

Potential applications:<br />

Comply with Policy on<br />

use of Animals:<br />

Proficient in reading,<br />

writing & speaking<br />

English:<br />

Benefits to<br />

individuals/institutions<br />

:<br />

Benefits to UK:<br />

none<br />

Not applicable<br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

Perception Group at INRIA Grenoble (where I am pursuing my Ph.D at the<br />

moment). The scope for collaboration is much higher at the later stages of the<br />

project, when we devise applications in motion capture, or in visual surveillance.<br />

Here, it is noteworthy to mention the highly successful industrial partnerships of<br />

the Vision group in Oxford Brookes. For example, the collaboration with the<br />

motion capture company Vicon has won the prestigious knowledge transfer<br />

partnership (KTP) in 2009.<br />

I have known <strong>Dr</strong>. <strong>Fabio</strong> <strong>Cuzzolin</strong> at INRIA Grenoble in France for 2 years, where<br />

we have participated together in seminars and lab-meetings. At INRIA, <strong>Dr</strong>.<br />

<strong>Cuzzolin</strong> and I have worked on similar problems (in 3D motion estimation), but<br />

from different perspectives. It provided us the opportunity to critique each other's<br />

ideas, and also to identify the individual strengths of our approaches. I wish to<br />

capitalize on that exchange, and collaborate more rigorously with <strong>Dr</strong>. <strong>Cuzzolin</strong>.<br />

Excellent reading, writing and spoken English capabilities. Received all<br />

professional education in English for the last 10 years. TOEFL 287/300<br />

Images and videos from visual sensors provide a wealth of information, potentially<br />

allowing more natural forms of interaction between humans and machines. This<br />

data, however, is bulky and expensive to store or transmit: its efficient but faithful<br />

representation could allow applications currently forbidden because of the scale of<br />

the data involved. Streamlining such interactions through cutting-edge large scale<br />

manifold learning will have enormous impact on productivity and as a<br />

consequence on economic growth. People will enjoy the effect of distributed<br />

computing on their lives, in the form of automated assistance to the elderly, or<br />

new home entertainment products not linked to inconvenient motion estimation<br />

devices. Companies active in entertainment, surveillance, and biometrics are all<br />

poised to receive clear benefits from such developments. Individuals' health will<br />

also benefit from advanced techniques of diagnosis based on semi-automatic<br />

recognition of diseases from 3D medical data.<br />

Clear benefits to the UK in terms of public security will come from the impact of<br />

the project on areas such as automatic surveillance. Biometrics such as face or<br />

iris recognition suffer from major limitations: they cannot be used at a distance,<br />

and require user cooperation, making them not practical in real-world scenarios.<br />

Methods based on gait analysis can be built on top of the proposed algorithm to<br />

design semi-automatic alert system with the potential of significantly improve the<br />

country's level of security.<br />

Besides, the scope of the techniques developed within the project is not limited to<br />

motion analysis or medical imaging. For instance, support to decision making and<br />

customization in business is another important area in which intelligent<br />

management of large amounts of data can have an impact. Using real time<br />

detailed data, products can be customized for individual clients, allowing<br />

companies to offer tailored services and boosting UK competitiveness.<br />

Page 7 of 8


<strong>Mr</strong> <strong>Kiran</strong> <strong>Varanasi</strong> / <strong>Dr</strong> <strong>Fabio</strong> <strong>Cuzzolin</strong> Newton International Fellowships - 2010<br />

Benefits to Overseas<br />

Country:<br />

Multidisciplinary<br />

Proposal:<br />

Financial Details<br />

Financial Details:<br />

Start Date:<br />

Duration (Years):<br />

Justification:<br />

Created: Friday, January 15, 2010 15:09 [Approved]<br />

The community’s awareness of the need for new, radical ways of learning from<br />

and making decisions based large scale data is arising. The Lisbon Strategy for<br />

growth and jobs explicitly mentions the need for Europe, and France in particular,<br />

to keep open to the most recent developments in ICT research. Failure to lead the<br />

mounting tide in large scale data management and learning, an issue at the root<br />

of most components of the connected society of the near future, would<br />

compromise the continent’s competitiveness and put serious obstacles before the<br />

realization of such scenarios.<br />

The potential of non-traditional manifold learning techniques in allowing efficient<br />

communication between different agents/sensors is apparent, making scenarios in<br />

which self-organizing systems (such as, for instance, self-organizing traffic lights)<br />

will adapt autonomously to changing requirements, reducing the need for human<br />

intervention or even centrally directed autonomous planning, realistic.<br />

Year Payment type Justification Amount Requested<br />

Year 1 Travel International 2,000.00<br />

Year 1 Other 0.00<br />

Year 1 Subsistence 24,000.00<br />

Year 1 Research Costs 8,000.00<br />

Year 2 Travel International 0.00<br />

Year 2 Other 0.00<br />

Year 2 Subsistence 24,000.00<br />

Year 2 Research Costs 8,000.00<br />

Total 66,000.00<br />

01/01/2011<br />

2<br />

£4000 per year for major international conferences (2 sums per year of £2000)<br />

£2000 per year for visits/seminars/collaborations travels in UK/Europe<br />

£2000 per year for equipment (eg cameras etc)<br />

Page 8 of 8

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!