I-CHASS|TU 7/29 Part II

Flickr Image: 

• Implications on preservation and reconstruction of Computer assisted Decision Processes
  • What can we reconstruct for the past?
  • Can we preserve the past?

• And how do you know the information granularity in order to help decision questions in the future.?....

This presentation made me reflect in a way that I had not before in all the man and computer power needed to preserve the past in order to protect the future. One of the main questions that stuck in my head was “is sampling an appropriate appraisal tool?” 

Bajcsy used examples from the Virtual Vellum to show on when comparing two manuscripts it is possible to detect “objects of interest”: gowns, clothes, crowns. Ine two of the illustration he exemplified (French manuscripts) the sequence of colors on a woman’s gown did not match on both documents. Bajcy’s challenge is: Can we design an algorithm that describes these features- gowns, crowns? There are a myriad of problems at the level of image analytics – in these French Illuminating manuscripts - such as the fact that the artists used different proportions of the face features. This is beautiful from the artistic and historical points of view but for the IT folks it is a problem – how to automate face registration? 

Another project that was described was “The papers of Abraham Lincoln”, a research initiative with the ultimate goal of making all writings by America’s 16th president available online. Bajcsy showed us a very complex diagram that I won’t even try to reproduce here – or even explain – on the start-to-end workflow challenges of the project including issues such as getting the documents from private owners to the museums. Bajcsy also said that we need to query the database in meaningful ways – he exemplified with the map of the girlfriends to whom Lincoln wrote letters – adding that this may or may not constitute a research question. It would be wonderful to get the project to the schools and they are hoping to get help from local schools in experimenting the environments with children but someone needs to maintain the database of course. 

Last but not least Bajcsy is developing 3D tele-immersive environments that are more affordable than the traditional ones - using visible spectrum and thermal infrared spectrum imaging for real-time robust 3D reconstruction of humans. He is interested in collaborative activities in these tele-immersive spaces. We saw a video of 2 dancers at Berkeley and UIUC dancing through the 3D space but physically in remote locations. It was interesting to see how the dances accommodate a technological flaw- the time lag – in the choreography in a very meaningful way. One other “flaw” – the fact that the bodies seemed to inter-penetrate each other was seen as a wonderful opportunity for choreographic exploration by the dancers. We were also show an application for wheelchair basketball coaching. 

Bajscy closed the morning by stressing problems that need to be addresses when developing the next generation o these 2D and 3D affordable environments such  as accuracy and speed calibration, the weakness of the original calibration, and the system robustness to illumination changes.

We then split among two labs where graduate students Rahul Malik, Suk Kyu Lee, and Hyu Jung Na showed us the tele-immersive environment setup. We could see the "composite" image projected on the 3 screens that composed the setup. The photos above this post document that experience.

In the afternoon we had the opportunity of hearing Bajcsy’s collaborators Michal Ondrejcek and Luigi Marini on respectively the Lincoln Papers and a demo of the cyber integrator, after which we were able to interact with the tool.  

A barbecue hosted Vernon Burton was next! Little did we know that we were to experience – besides a great barbecue!!! -  the sudden climate changes of Illinois. At 7PM, a heavy rain made everyone run inside Vernon’s home for more pork and salmon - or go home because it had been a long day and next day’s program was excellent (too!)