Cecilia Farias
Cecilia worked with Nora Tramm in Professor Biron's lab on a project in image recognition. The lab's job is to track the behavior of primitive worms whose neurons can be individually modified. We know about the worm's behavior from its shape and its movements. We record these over time with a video camera. Then the video record needs to be analyzed in order to codify the behavior that is occuring. For this one must accurately reconstruct the worm's shape. The lab has developed a software program to accomplish this reconstruction. But the software makes mistakes sometimes. For example it often makes errors in the shape of the head, and this head shape is important in determining the physiological state of the worm.
Cecilia's project was to test a new software package for determining the shape of the worm. She analyzed many video frames by both methods, noting how the two methods compared using various criteria. She found that the new program was often more accurate than the old one, as was hoped. However, much of this improved accuracy proved to be attainable by modifying the old program. More importantly, the new program sometimes made unacceptable errors, such as identifying features of the background as being part of the worm. Cecilia is now finishing the project by exploring how the performance of the software can be improved by preprocessing the video images that are given to the software using operations like contrast enhancement and background removal. Right hand picture shows successive stages of preprocessing.
---TW 5 March 2012, OK'd by Nora
8Mar
Javier Baeza
A water drop on a hot metal plate begins to dance, driven by the steam it gives off. Javier Baeza, working with postdoc Justin Burton in Prof. Nagel's lab, is trying to see the steps of this dance. One part of the dance is jumping: the drops jump bodily off the plate, and they remain in the air long enough to oscillate back and forth. Another part of the dance is twirling. Big drops distort from a round shape into a bumpy shape, and the bumps move around, as shown at right.. In both of these motions natural oscillations of the drop get amplified by the positive feedback, bouncing pushes the drop against the plate; this creates more steam and this gives a kick that increases the bouncing. The pictures show Javier with his fast video camera and one of its video frames showing a drop distorted into a rotating lozenge shape. The project is proceding after Javier's departure. The group will study the frequency spectrum with relation to the capillary oscillations, based on Javier's preliminary measurements. They will extend the interferommetric method developed by intern Andrès Franco in 2011.
----TW 29February 2012
César Parra
Cesar Parra studied the novel flocking phenomenon seen in suspensions of bacteria such as Bacillus Subtilis. His advisor was Dr. Igor Aronson of Argonne National Lab. Bacteria swimming in random directions become aligned by their interactions, as shown in the picture at left [from A. Sokolov et al. Phys Rev. Lett 98, 158102 (2007)]. The alignment may result from simple collisions or from the flow around each bacterium caused by its swimming. Igor and Cesar aimed to explain the abrupt transition from unaligned to aligned as one gradually increases, eg, the density. They explored a simple mechanism using the idea that alignment must be enhanced if the orientational diffusion that randomizes the orientation is reduced. They aimed to see how much this diffusion would be reduced when the concentration was slightly increased. They suspect that the main effects seem to arise from the spatial fluctuations of concentration rather than from the average concentration. Cesar and Dr. Aronson are continuing to work together; a publication is in progress.
---TW 24February 2012, OK'ed by
Cesar June 7