It has been exactly one month since I started taking 20 photos a day. Although this commitment barely maintains an element of joy, it persists and I would love to see how far it can go. Why am I collecting all of this data? Being in the field of computer vision (my daytime occupation), I estimate that within a few decades we will be able to index visual data and extract rich information from it, including information about one’s daily life.

A computer will have a reliable understanding of image contents — understanding that surpasses the human eye and mind. Want a photo that which a Labrador drinking water from a fountain on a sunny day? Want a descriptive verbal interpretation of a given distorted image? This will probably be practical in the distant future. It is a machine learning/pattern recognition task taken to extremity.

The Maths Tower on campus (undergoing demolition)

I am going to present recent research work at UCL tomorrow. Bound to this presentation are some graphics files (with captions). Older, unused slides are available to anyone who is curious. I will most likely modify text on the train and upload the changes during the weekend.

To test my algorithm for a paper (see related work), I needed to generate warped face images (approximately 300 of them) and found myself with a well-occupied workspace. I use the GIMP animation Script FU called “Ripple” to achieve the necessary effects. The script takes an input image and outputs a sequence of images which make up a video of a ripple effect, like a flag.

Example input:

One image of a sequence of outputs:

Thank you, Adam D. Moss, for sharing that excellent script.

One of the more amusing stories from Slashdot talks about an academic paper in computer science, which was accepted to be presented at a conference. The catch: all text, figures and references were randomly-generated and are completely incoherent (see Web site). Will a randomly-generated talk be delivered as the guys at MIT promise?

will apparently deliver a talk at University College London next week. Here is a gist of my current research work:

We are given a collection of raw images depicting an object such as the brain. We then construct a model which can produce synthetic images as shown below.

We developed a framework that can quantitatively evaluate such models, having constructed them automatically. For more details, see a previous entry on the subject or confer my research pages.

This work has recently been accepted for presentation in IPMI, which will be hosted in Colorado. I currently work on papers for MICCAI (due later tonight) in Palm Springs, California and BMVC in Oxford.

have the pleasure to have been ranked second in the world for my contributions to MATLAB last month. This equates to my position in July 2004, but exceeds the number of monthly downloads, which was just under 2,000 at the time. To celebrate this occasion, I treated myself to a cup of coffee with two three sugar.

This photo (direct link) has got to be seen for its enormous size to be comprehended. Zooming appears almost infinite, without loss in quality. The site provides the tools for viewing and interacting with the data in real time.

If this photo were printed, it would measure 6.67 m by 2.67 m (300 dpi). The photograph shows Delft and its surroundings in the autumn of 2004.

On the right: my illustration of the extent of magnification