Wednesday, January 20, 2010

The perception of scenes with natural light

This Monday, on January 18th, 2010, I came to another neuroscience colloquium in Cognium. The talk is presented by Prof. Dr. Tom Troscianko from the Department of Experimental Psychology, University of Bristol, UK.

The title of the talk is "The perception of scenes with natural light".

The talk is about the visual perception of animals (Homo Sapiens included) as they see some scenes with natural light, in this case is sun light.

The first part of the talk is about the visual perception of primates and some birds: How primate see fruit, how about birds. Some models are shown in the slides. What happened with the monochromatic animals (you can say color blind)? Most primates are not color blind so they can see the contrast better than the birds. The first part is mainly about the perception of colors and contrast. Pictures and their histogram were shown in the talk.

The second part is about the effect of shadow. Can human perceive shadow from natural light and from manipulated image? And how fast?
The first experiment is shadow direction. There are pictures of standing cylinders and their shadow with the "light source" from above. Which shadow has wrong direction?
Then the picture is rotated upside down. Which shadow has wrong direction?
Eye movement and time are measured.
It happened that humans are faster with the light source from below. Human is not aware with the light source from above, for example sun light.

The third part is about the estethic perception. In this part, the talk is about the visual perception of sunset. Is sunset beautiful? Why do people like sunset? Can we measure the beauty of the sunset?
Google give more than 40 millions results for sunset and more than 30 millions results for sunset pictures.
There are also experiments on a ship. People gathers more on the side of a ship where they can see a sunset.
There are two conditions in sunset. When the sun is high, there is Rayleigh scattering of sunlight. The sky looks red at that time. When the sun is low, very near horizon, there is Mie Scattering of sunlight. The sky looks blue with the horizon looks red and yellow.
The experiment is conducted the eye movement.
The result shows that when the sun is high, human prefer to see the sun in the sunset, not the halo or the red sky. On the contrary, after the sun reaches the horizon, human prefer to see the halo and the sky. From this experiment, the "beauty" of the sunset can be measured.

So the whole talk is about the perception of color and contrast, shadow and the esthetic. We can know the research interest of Prof. Troscianko from here.


Next talk will be interesting. It will be on February 1st, 2010. It is about neuroethic. The title will be "Von der Neuroethik zur Bewusstseinsethik". Well, reading "mind" can lead to some ethical problems. I will make another blog post of the next talk.

Improvement of Response Times in SSVEP-based Brain-Computer Interface

Starting on December 10th, 2009, I have a master thesis. The thesis should be submitted on May 27th, 2010. The presentation will be conducted in June (I hope).

The title of thesis is "Improvement of Response Times in SSVEP-based Brain-Computer Interface".

The supervisors are
  • Prof. Dr.-Ing Axel Gräser
  • Dr.-Ing. Ivan Volosyak
  • Thorsten Lüth, Dipl.-Ing
The thesis is conducted in the Institute of Automation (IAT) at the University of Bremen.
The research group is no longer called BrainRobot. The name is now BRAIN, which stands for Brain-computer interfaces with Rapid Automated Interfaces for Nonexperts.
Yes, it is funded by European Union. We want to keep up with all research groups in the USA (and Canada) and in the Asia Pasific (China, Japan, Korea, etc).

Back to my thesis!
The proposed question behind the thesis is whether we can improve response times of our system in detecting SSVEP patterns from a subject. We can say that I want to make Bremen BCI system (a little bit) faster than before. I am using time-series manipulation algorithm to do so.

More details will be told in other blog posts.