Claudio Caraffi Email & Phone Number

Berlin, De

Berlin, De

After the exciting experience in Silicon Valley, I took the opportunity to spend a sabbatical in my home town after 10 years abroad, dedicating time to my family and sports.Then we came back to Berlin where I resumed my career.

Menlo Park, Ca, Us

Software development for image processing for optimized eye feature extraction, user identification, result propagation and visualization, and integration in a compex multi-sensor environment.

Cupertino, California, Us

Preparation, transition and technology hand over in the context of SMI acquisition by Apple

(Acquired by Apple in June 2017)Development of flexible eye-tracking algorithms and software for platforms ranging from desktop to virtual and augmented reality systems.User’s gaze reconstruction requires quick and precise (sub-pixel) elaboration of eye features. Challengescome from variability of users’ appearance, posture and eye characteristics, of the environmental conditions (e.g. disrupting reflections), and of the hardware, especially seen in a number of customizations for roughly all most known Virtual Reality devices. In these years we have been working 360 degrees on these aspects. I established with my colleagues a new concept for massive evaluation of implemented methods. I have redesigned the pupil and iris detection and reconstruction algorithm, and contributed to the reshaping and optimization of the other parts of the pipeline.Side activities:– Organized image acquisition campaign– Patent conception and writing (author and co-author of 2 submissions)– Development and maintenance of multi-platform CMake building system (Windows, Linux, Android -llvm, gcc compiler)– CPU and DSP vectorization of code– Job candidates selection and evaluation

Brussels, Be

Responsible for the creation and development of new concepts for traffic scene perception and risk detection to be applied in the field of automotive safety:- Manage cooperations with different research partners for algorithm development, setting scientifictargets and work plans.- Introduced new concepts for risk representation, early detection, persistent tracking.- Created a modular software platform for algorithm development and integration.- Introduced an innovative method for semi-automatic vehicle annotation and algorithm evaluation as alternative to expensive manual annotation, dataset and evaluation source code released (See projects below).- Responsible for the set-up of a multi-sensor data acquisition vehicle (11 cameras for a 360 fieldof view, 1 Velodyne laser-scanner, PosLV, CAN acquisition), revealed to the scientific community at ECCV2012. Coordination of several involved divisions.- Periodical presentations of research work to colleagues and board members in internal exhibition and global meetings (US, Japan), outstanding feedback.- In the summer 2012, in charge of transferring and testing 2 Computer Vision algorithms (vehicledetection and tracking and lane detection) in the autonomous vehicle prototype developed inToyota America (revealed at CES 2013); successful integration with autonomous driving during aperiod of 10 weeks spent in Ann Arbor (MI) cooperating in a multi-cultural team.My interest is in low computational cost algorithms, in particular using a priori considerations to exclude the impossible and to concentrate the processing power on key information subsets. I privilege monocamera algorithms, and integration with other cheap source of information, targeting low cost solution.

Toyota City, Aichi Prefecture, Jp

Short term contract which lead to my recruitment.

Computer vision applied to intelligent vehicles. Lane and traffic signs detection.

Perception systems using Computer Vision for manned and unmanned vehicles in on-road and off-road environments. My research regarded problems related to system calibration,image stabilization, stereo off-road obstacle detection, road slope profile estimation,lane perception, and (in cooperation with Magneti Marelli Holding S.p.A.) traffic-sign recognition.Participation to all edition of DARPA Autonomous Vehicle Challenge (2004, 2005, 2007).The systems I developed were a fundamental factor first for the set-up of our direct cooperation with Oshkosh Truck Corporation in 2004,and second, given our success, of the renewed cooperation in 2006 for the development of 5 vision systems involving more than 10 research engineers from VisLab.These projects have been very important contribution for the growth, also in prestige, of VisLab.

Oshkosh, Wisconsin, Us

Developer and responsible of vision perception systems for autonomous vehicles TerraMax andTerraMax2, in an international team working environment.In the period September 2003 – February 2006 I have been developer and responsible of the visionsystem (off-road stereo images stabilization and obstacle detection) of the TerraMax autonomous vehicle for 2004 and 2005 edition of DARPA Grand Challenge. In October 2005, the TerraMax vehicle has been one of only 5 to complete the race, the only one using vision as main sensor.I contributed mounting the same system in a second autonomous vehicle for a demonstration in Yuma, AZ.In the period June 2006 – November 2007 I have been developer and responsible of the vision system (short and long range lane detection) of the TerraMax2 autonomous vehicle for 2007 DARPA Urban Challenge.

Cedar Rapids, Iowa, Us

In the period November 2004 - February 2006 I have collaborated with Rockwell Collins as developer of the vision system (image stabilization and obstacle detection) of the TerraMax autonomous vehicle that finished 2005 DARPA Grand Challenge, and contributing mounting the same system in a second autonomous vehicle for a test which took place in Yuma, AZ

Phd, Computer Vision

Computer Vision

Msc, Computer Science