I'm an Artist, Engineer and a Musician
I seek for the boundaries between imaginary and real;
trying to bridge between human and machine
I made some prints, played some chords
I solve problems for a living, probably using code;
I took some pictures, made some gifs,
The things I make are mostly related to:
android arduino camera sensors sonification servo video installation hackathon kinect photography code vision gifs hardware electronics fun 3d movement processing music
interaction photography code vision gifs hardware electronics fun 3d movement processing music android interaction photography code vision gifs hardware electronics fun 3d movement processing arduino camera sensors sonification servo video installation
video installation hackathon kinect interaction photography code vision gifs photography code vision gifs hardware electronics fun 3d movement processing music electronics fun 3d
movement processing music android arduino camera sensors sonification servo video installation hackathon kinect interaction photography code vision gifs hardware movement processing music android arduino camera sensors sonification servo
code vision gifs hardware electronics fun 3d movement processing music android arduino camera sensors sonification servo video installation hackathon kinstallation hackathon kinect interaction photography code vision gifs hardware inect interaction
I made this website :)
I programmed animations for Reuven Berman Kadim
I made a Image Filter for Canva
I got an Honoraria grant to build a Huge Chameleon Sculpture for Burning-Man
I wrote a compiler in 3 days & 50 lines of code
I built an Eye-control interface for people with ALS
I developed Demos and Mockups for natural gesture TV interfaces
I developed Smart-Camera Android App for Optimizing Image Composition
I built an Experiment to analyze people behavior with AI
I made an on/off switch using pants zipper
I made some prints using various techniques
I helped this person build a Camera Focus Controller
I helped this dude build an Gesture Controlled Orchestra
I helped this guy build an Accordion Controller
I made a Visualisation of Wikipedia articles
I developed and taught a Workshop for making motion-games using Depth Camera and Unity
I taught a Workshop for making motion-games using Depth Camera and Unity
I am currently self-employed doing freelance work
I got an Honoraria grant to build a Huge Sculpture for Burning
Man 2018
The “Come-Million!” is a massive Chameleon sculpture built for Burning Man 2018 in Black Rock
City,
Nevada.
The sculpture received an Honoraria grant and was built by a collective of artists, dreamers, engineers
and various of people from the local burner community.
It is 20 Meter long and 7 meters high. It is covered with 20,000 LEDs that sit on a
crafted
steel-frame.
It offers a delightful, colorful animations full of light, that reacts and respond to its surroundings.
Writing a compiler means to understand the language itself - quite mind boggling;
rarely you need to know how Functions work, or need to define the control-flow itself.
I am working on a mini tutorial to show how FALSE can be built starting from the concept of calculator,
simply adding numbers - and extending it with words - variables and functions, up to a full blown compiler.
Using a live video stream from a camera located on glasses frame, pointed directly towards user's eye.
The video is processed and analyzed using computer vision algorithm to track the eye movements
and translates them to keystrokes (up/down/left/right)
inspired by the eyeWriter project, and uses same low-cost
hardware (webcam on plastic glasses).
What makes EyeNav different is that unlike direct, mouse-like cursor manipulation, which can be tiresome
-
it is based on discrete gestures (like keyboard), for navigation.
The project was conceived, born, and took its first steps in 12 hours, during a Hackathon for development of
apps for disabled people
I developed Demos and Mockups for natural gesture TV interfaces at Primesense
Reach is Primesense's TV control paradigm, that demonstrates a concept of direct media access, using natural
gestures. Developed at Primesense UX Group.
I developed "Smart-Camera" Android App for Optimizing Image Composition
Recent years have seen dramatic progress both in the hardware and software driving digital photograpy. We
now
can utilize the processing power of smartphones to create stunning visual effects.
Alas, almost no effort was made to improve the _composition_ of the images - while shooting the photo. This
work
addresses the challenge by combining computer-vision and photography-theory to suggest an automatic
'ViewFinder'.
‘We Are Family’ is an interactive, typographic-animated-gif installation that make use of fragments from
found
online political, brutal, pornographic and other videos to create a new kind of typographic visual language.
Try it yourself at project page
Read more at Oded Ezer website
ZipperSwitch
I made an on/off switch using pants zipper
A proof of concept for using a zipper as an on/off switch:
The switch triggers a tactile feedback through servo hidden in the belt,
to send the wearer a silent and kindly reminder - if the zipper is left open for too long.
An experimental research in machine-learning that identifies and analyzes the concept of social normalcy.
The person selected is examined by the machine and is added to its algorithmically constructed image
database.
Part of several exhibitions shown in Israel, Athens and Kiev
I wrote a Patented! gradient descent search algorithm for wireless power charging solution
The base station self adjusts to the user's device
The algorithm was written in C, firmware using under 64 Bytes of RAM
Video Accordion
I helped this guy build an
Accordion Video Controller
An accordion that controls the playback of the video, where each key picks a different performance-clip to
play,
while the direction and speed of the video playback is controlled by the accordion's bellows.
Watch the whole circus in vimeo
WikiWars
I made a Visualisation of Wikipedia articles
Using advanced dimensionality reduction and text analysis techniques in order to create a two dimensional
semantic map of Wikipedia terms. Each term was represented by an individual pixel color-coded by how
contentious
the topic was.
The “controversiality measure” is extracted by carefully analyzing the revision history of each term in
order to identify combative erasing and rewriting, indicative of a highly disputed topics.
Dance Jockey
Movie Weaver
I made a Markov-Model Automatic Video Editor
"Movie Weaver" is an automatic video editor that utilizes the power of probabilistic models along with
filmmaking paradigms to turn a collection of video clips into a movie.
Note the smooth color transition of a optimized sequence of clips:
compared to a random sequence:
For more details read the book or see
the
poster
Talks, Events
Workshops / Hackathons
I developed and taught a Workshop for making motion-games using Depth Camera and Unity
Watch a show reel the results of 5-hours session:
The workshop was held at:
Co-organized of "INFECTED" Hackathon at
the "Center of Design and Technology".
Mentor and technical assitant for teams who developed projects with Primesense 3D camera
Talks
Speaker at GDC, Game Developer Conference 2012, Presenting "How to create
Motion-game from scratch in 30 minutes"
Speaker at [TAU film festival](https://www.taufilmfest.com) 14th International Student Film Festival, Tel-Aviv.
Presenting at the 'New Media Conference' Primesense's 3D imaging technology and it possible usage in the film
industry.
Résumé / Curriculum Vitae
Work Experience
I'm a Software Engineer with 15+ years of hands on experience.
I have M.Sc and B.Sc in Computer Science and Computer Engineering from the Hebrew University of
Jerusalem,
Israel.
My research focus was on Image Processing and Numerical Analysis algorithms for Computer Vision
applications
I have built live demos for start-up companies and site specific art installations for museums for fun
and
profit
I have experience working with various technologies, frameworks and tools for real-time systems on
native
and web platforms
I worked for several startup companies:
Powermat, who developed the first wireless charging solution in the world; and
PrimeSense who invented the Kinect, and sold to Apple.
I was one of the first employees for NextDroid, a robotics startup in Boston who grew from 5
people
more than 100
and an part of Snapchat special applications team
I am currently self-employed doing freelance work.
2019 - Now:
Software Engineer
Self-Employed
AR navigation app
Proof of Concept AR Fitness,
full-body 3D tracking, exercise detection
demo: counting squats
AR navigation app
2017 - 2018:
Software Engineer
Global Brand Team
@ Snapchat
Designed and engineered live installations for worldwide events
2015 - 2017:
Software Lead
Advanced Robotics R&D
@ NextDroid
Planned, initialized and executed projects for autonomous robotic vehicles
Integrated computer vision, path planning and navigation
2009 - 2011:
Firmware Engineer
R&D group
@ Powermat
Lead developer of "Powermat Anywhere", a wireless power charging solution, self adjusts using
gradient descent search algorithm using less than 64 bytes of RAM
Patent
#9,478,991
System and method for transferring power inductively over an extended region
An inductive power transfer system operable in a plurality of modes comprising an inductive
power
transmitter capable of providing power to the inductive power receiver over an extended region.
The
system may be switchable between the various modes by means of a mode selector operable to
activate
various features as required, such as: an alignment mechanism a resonance tuner, an auxiliary
coil
arrangement or a resonance seeking arrangement. Associated methods are taught.
Patent
#9,459,758
Gesture-based interface with enhanced features
A method includes presenting, on a display coupled to a computer, an image of a keyboard comprising
multiple keys, and receiving a sequence of three-dimensional (3D) maps including a hand of a user
positioned in proximity to the display. An initial portion of the sequence of 3D maps is processed
to detect a transverse gesture performed by a hand of a user positioned in proximity to the display,
and a cursor is presented on the display at a position indicated by the transverse gesture. While
presenting the cursor in proximity to the one of the multiple keys, one of the multiple keys is
selected upon detecting a grab gesture followed by a pull gesture followed by a release gesture in a
subsequent portion of the sequence of 3D maps.
Patent
#9,377,865
Zoom-based gesture user interface
A method includes arranging, by a computer, multiple interactive objects as a hierarchical
data
structure, each node of the hierarchical data structure associated with a respective one of
the
multiple interactive objects, and presenting, on a display coupled to the computer, a first
subset
of the multiple interactive objects that are associated with one or more child nodes of one
of
the
multiple interactive objects. A sequence of three-dimensional (3D) maps including at least
part
of a
hand of a user positioned in proximity to the display is received, and the hand performing a
transverse gesture followed by a grab gesture followed by a longitudinal gesture followed by
an
execute gesture is identified in the sequence of three-dimensional (3D) maps, and an
operation
associated with the selected object is accordingly performed.
Reach is Primesense's TV control paradigm, that demonstrates a concept of direct media access, using natural
gestures. Developed at Primesense UX Group.
A proof of concept for using a zipper as an on/off switch:
The switch triggers a tactile feedback through servo hidden in the belt,
to send the wearer a silent and kindly reminder - if the zipper is left open for too long.
I wrote a 
The base station self adjusts to the user's device
The algorithm was written in C, firmware using under 64 Bytes of RAM
compared to a random sequence:
For more details read the 
Recorded album of songs and covers on one night at Studio-one, July 20, 2004
