RARELY find the time or even the need to write about my research in this blog, either the Linux- and software-related research or my job where I study geometric methods and develop new techniques that nobody tried before. The reason for inaction regarding the latter is that I cumulatively prepare a large report (over 300 pages so far) which will be published in full in this Web site when it’s all polished, along with code. I posted some material before. For the former, there is a site dedicated to that and it is called Techrights. That site took many years to build a reputation and PageRank 7. It now attracts a huge number of visitors and serves as my main blogging platform (I have about 10 installed WordPress blogs in total). Schestowitz.com became a dumping ground for random thoughts and stories that do not fit elsewhere. Thanks to all those who follow this blog in spite of disorder.

For more regular updates from me, please also follow Techrights. It is more polished than this Web site.

O, TODAY was a little dull. In the midst of boredom scripts were produced to turn a sequence of images into an animation of small size with an example of crop, border, and resize (ImageMagick required). Let’s go through this one step at a time. Put all the images from the sequence in a working directory.

mogrify -resize 25% -border 3 *

This adds a border and also resizes all images at the same time.

mogrify -crop 300x70+0+40 *

This one crops the images in accordance with the coordinates given (width, height, horizontal offset, vertical offset). Once everything is processed:

animate *

Save the resultant animation as GIF, for example (there are other options). Here is a sample:

That specific animation provides insight into variability of surfaces on which geodesic measures are taken. For particular cases, detection is made harder by motion around the eyes (including closure), but this is one of the caveats of dealing with surfaces of these kinds. It would be easier with biomedical data.

HE FOLLOWING research notes have just been uploaded, hopefully to help others understand the work undertaken and what ideas are being tested (generalised to the extent of being applicable to any type of data, including biomedical). The 200+ pages long document which is more complete is still work in progress and it will be published at a later date. Thanks to the ERC for funding this work.

INALLY, after nearly a week since taking and returning from a break, I received the server I have been waiting for. It arrived on Monday afternoon and today I finished configuring it, running alongside the triple-head workstation which I will use to do computer vision jobs and other development/admin work. The research requires a lot of screen space and sometimes the RAM/CPUload goes rather high, which helps not at all unless there is a second motherboard at hand (and a fallback machine.

The server and the stereo beneath the lefthand screen

Frontal view

Over the shoulder

t various stages throughout my career (I am 29 now), myself and others pondered starting a blog about Open Access, Open Data or open research (opening one’s lab, including data, methods, etc.), but since I already spend more time than I can afford advocating GNU/Linux, I ended up spending a lot of my energy fighting against software patents. This is one aspect among several involving the sharing and transparency of the sciences. There is of course also a dimension associated with copyrights and despite its importance I no longer have time to pursue the issue; many people already do so anyway. To put the core argument succinctly, nowadays when we have the Internet and we have a presentation layer such as the World Wide Web (for exchange of inter-connected rich media), we no longer depend on many analogue technologies and the notion of scarcity must cope and catch up with what’s available. To expect people to publish their findings only in paper form with pagination is to pretend that we are locked into legacy, which we should not. Computer Vision is considerably limited by paper. There are much better and faster ways of getting messages across, especially in this age of information overload. Moreover, travelling for presentation to a crowd (conference) is less necessary now that many people’s Internet connections permit video streaming at a good resolution. Romanticising over the nostalgia which is 20^th century-esque research is no longer beneficial as it leads to inheriting inherent limitations. I was reminded of this in an IRC conversation last night. It was agreed upon that the theory about publishers exploiting academics to write, review and even edit entire publications for the publishers to profit from at the expense of those academics is something that needs to be stopped. Many academics these days have their own homepages and sometimes their blogs too. They can publish a lot of material there and let the quality/accuracy be determined by citation, e.g. something like the PageRank system which merely inherits the ideas of algorithms before its time (take away Google’s/Stanford’s patent while at it). The issue of course is that people cannot reference papers by Internet addresses, at least not by conventional means. The idea that papers should be accessible through libraries though is outdated as libraries too are going somewhat extinct and the speed of working there is inferior. Other than self pride and honour from peers, what incentive is there really to being heavily involved in the publication industry which benefits publishers and offers writers not so much exposure anymore? Fewer people seem to be searching journals; they use external search engines or Wikipedia instead (it’s multi-lingual). The debate becomes ever more relevant now that Aaron Schwartz is being hounded for just doing his job.

Eventually, prices will converge somewhere around zero not because academic work is worthless but because people’s goal is to share and disseminate their ideas, not to serve someone else’s paywall. The real worth of research is derived from its reach, not its scarcity.

s a demonstration of canonical forms and stress reduction complemented/guided by multidimesional scaling, I’ve created an animation which shows the process applied to each image in the Face Recognition Grand Challenge (FRGC) 2.0 set — albeit Fall Semester only in this case — in turn, in order to approach a more mutually-isometric and pose-agnostic state where distances are tied to inherent surface details (curvature, size, etc.) and the accompanying static image, as seen below, shows the original image too (added at the top). To use this within an objective function it will need to be clearer how points are selected consistently and where correspondences can autonomously be chosen to improve overall performance. The triangulation in this case is Delaunay-based although 3 methods have been implemented and they offer room for further experimental work. The factors affecting performance may be the PCA component, the triangulation, the placement of points, the optimisation of lengths, the pre-processing (ICP for instance), and few minor technicalities less worthy of consideration. Each one of these represents one parameter among many but feasibility tests — those exploring whether the overall framework is effective in the first place (distances as an encoded signature resistant to expressions) — must come first. Based on a preliminary look, this ought to serve as a reasonable discriminant, but many of the pertinent parts of the framework may need tweaking based on trials and errors.

Note: this is an except from an ongoing project and a document exceeding 200 pages so far. It will be released later this year.

his series of videos is a rough explanation of the approach taken in order to utilise principal component analysts (PCA) in the task of shape classification. This was done without preparation or second takes, so the quality and clarity are not particularly high.

PCA and Multidimensional Scaling in Shapes – Part 0

PCA and Multidimensional Scaling in Shapes – Part 1

PCA and Multidimensional Scaling in Shapes – Part 2

PCA and Multidimensional Scaling in Shapes – Part 3

PCA and Multidimensional Scaling in Shapes – Part 4

PCA and Multidimensional Scaling in Shapes – Part 5

PCA and Multidimensional Scaling in Shapes – Part 6

PCA and Multidimensional Scaling in Shapes – Part 7

PCA and Multidimensional Scaling in Shapes – Part 8

PCA and Multidimensional Scaling in Shapes – Part 9

PCA and Multidimensional Scaling in Shapes – Part 10

PCA and Multidimensional Scaling in Shapes – Part 11

PCA and Multidimensional Scaling in Shapes – Part 12

PCA and Multidimensional Scaling in Shapes – Part 13

PCA and Multidimensional Scaling in Shapes – Part 14

PCA and Multidimensional Scaling in Shapes – Part 15