Bouncing Balls

A phase-space plot of height v. exaggeration for two different types of balls with two different coefficients of restitution.

Nothing important here, just a fun looking plot.





Fencing Eyetracking Data

A while back, we did a nice study on eye movements during fencing.

This is a figure from one of the posters, we should finish this and publish it some day. Should do a lot of things.




Medical Imaging, 1986 Style.

I was going through some slides for a talk today and I stumbled on this one from 1986 of a tool I wrote for analyzing cardiac ejection fraction.

You can tell it’s old because it’s in black and white.

It was written on a Sun II named jabba (I left the prompt at the top to show you) at the Computer Graphics Research Group at The Ohio State University. It used a client-server setup where the heavy duty image processing was done on a Convex mini-supercomputer (where ‘mini’ = the size of a small car) and the user interface was done on the Sun. Indeed, it used a ‘front end’ and ‘kernel’ model that Theo Grey and Stephen Wolfram adopted (not in any way due to my stuff, I’m certain) for Mathematica. Coincidentally I would go on to beta test early versions of Mathematica, befriend Stephen and Theo, and even work for WRI for a while.

I wrote a paper about this model, submitted it to the Convex User Group and received a Macintosh SE as a prize for it! (Scott Dyer, also of CGRG, wrote about a rendering engine they were producing and also received an Mac.) Nowdays, when I submit a paper, I wait for 3 months for it to be ‘peer reviewed’, make revisions to suit some reviewer or editor’s agenda and then pay for the privilege of it being placed in a ‘blessed publication (aka web server)’. Amen. Praise science. (This subject is a whole blog post in itself, of course.)

Part of this tool (the cine module) apparently remained the ‘go to’ play-back tool at CGRG for several years after I left. After I wrote this, I left and went to Pixar where I designed some medical imaging software user interfaces. This was in the stone ages of UX/UI, no one called it those things, (‘GUI’ maybe), we called it a ‘user in your face’.

Man we were clever.

Mmmm, fresh data

Mmmm, fresh data

Travels in the uncanny valley

Here’s a screencast of my talk at the Rochester Institute of Technology Symposium on VR/AR Technology.

It is from 3 December and a little low-key, so drink a little coffee before you start watching so you don’t fall asleep. Amazingly, the MAGIC Center had no lecture recording, so I’ve recreated it here with about 60% less energy because I was sitting down the whole time.



In 1970, Masahiro Mori posited the existence of an “uncanny valley” in the human perception and appreciation of robotics. This hypothesis suggests that, as robots become more ‘human like’ that our appreciation of them increases — right up to a crucial point where we find them too realistic but not perfectly so. At this point, human preference plummets into the aforementioned chasm. Despite the fact that much has been written about this phenomena, there is little scientific evidence of its existence. Obviously this phenomena isn’t limited to robotics and has been observed in many other areas. These include the fine arts, especially in photorealistic painting, sculpture, computer graphics, traditional animation, and more recently in the rotoscoping / motion capture used in contemporary filmmaking. The informal and heuristic practices of the fine arts, especially those of traditional animation have much to offer to our understanding of the appearance of reality. In this talk, we will examine some historical perspective of the UV as well as virtual and augmented reality-centric questions.

Enhancing research with Plenary Labs

A new pub with some good friends on the impending doom of science funding and what to do about it. Authored well before the new administration was a gleam in our collective eyes.

Citation: Pawan Sinha, Peter Bex, Margaret Kjelgaard, Flip Phillips; Enhancing research with Plenary Labs. Sci Public Policy 2016 scw051. doi: 10.1093/scipol/scw051


The two most evident crises in the advanced research ecosystem in the USA are scarcity of funds and scarcity of jobs. We argue that both of these are outcomes of a flawed resource usage model and propose an alternative approach that can help alleviate these challenges. Named ‘Plenary Labs’, this approach is designed to act as a counterpoint to the traditional schema wherein each laboratory acts as a self-contained silo with a full complement of equipment and personnel to advance the principal investigator’s research program. This schema results in redundancy across labs, as well as an inflated need for research assistants. Plenary Labs ameliorate both of these issues by consolidating equipment and technical manpower. By democratizing access to cutting edge resources, reducing the time and costs involved in experimental research, and reducing the imbalance between supply and demand for jobs, Plenary Labs have the potential to significantly enhance research.


Free Access: Grab a delightful open access copy here.

Quick class description

For those of you looking for the course description of “Vision in Animals and Machines” that I failed to send to the Registrar, look no more!

Vision in Animals and Machines

PS-312, Fall 2017

This class will focus on ‘vision’ in its many forms — ranging from single-celled animals to complicated ‘artificial intelligence’ based computer vision systems. We will survey the biological and evolutionary processes that led to our human sense of sight as well as how both natural and artificial systems have influenced our understanding of ‘vision.’ We will explore computational models of human vision as well as current state-of-the-art deep learning techniques for image and scene identification. Ultimately we will settle once-and-for-all whether ‘vision’ as practiced by machines bears any resemblance at all to the behavior of ‘vision’ in natural systems.

Contact for information.

Fuzzy Logic package for Mathematica

I was working on a project in Gießen last week and needed some code I wrote back in the late 1980’s / early 1990’s.

Sadly, this is no replacement for the super cool FuzzyLogic Mathematica package (RIP).

It is simply the ten fuzzy boolean operators + a few things here and there that add to the fun. It implements the traditional ‘Zadeh’ operators, along with Yager and HyperbolicParaboloid variants, selectable with Method.

Let’s hear it for being a complete packrat. Added to the Software, Data & Geometry page is a very old fuzzy logic package. You can make a direct hop over to GitHub here.

More drawing research

Flip Phillips is Protected by Akismet | Powered by WordPress