User:Brian P. Josey/Notebook/2010/07/29
|Project name|| Main project page|
Previous entry Next entry
Tracking Software with Larry's Help
I've been having a little trouble working on the tracking software, and unfortunately motivating myself to tackle it also. So I asked Larry for help, and he helped me this morning begin to set up the tracking software. The first thing that he set out to do was to establish a way for the computer to tell were the droplets of water were in the oil emulsion. While it is easy to for us to see where the droplets are, picking them out and track them with our eyes, it is more difficult for the computer to do that. So what Larry did was to find the peaks of brightness that occur at the center of the droplets, and isolate them. Then he removed all the background noise, and drew circles and boxes around the bright parts. These define the points that we are interested in and will be tracked by the computer.
One issue that did come up was how to distinguish a single droplet when it is tagged multiple times by the computer. This happened for one particular droplet, that had a textured pattern and looked like a crescent when the high points were picked out. While we could fill in the region of interest for other large droplets that generated an empty circle, we couldn't apply this to the crescent. So it would be counted multiple times, or misrepresent the data from the droplet. Regardless of these difficulties, we were able to successfully tag plenty of regions of interest on one of the images, and it is my goal to apply the same process to the whole sequence of images.
Finding the Radius of a Droplet
My emulsion experiment hinges on finding two points of data for each ferritin solution droplet in the emulsion, it's radius and the velocity the droplet travels in the flow cell. From the radius, I can determine the total number of ferritin in a droplet, and I can find the net magnetic force from the velocity. Together, this will give me the average magnetic force per ferritin protein, and either support, or hinder my theory. While finding the velocity and force is fairly simple as soon as I finish the tracking software, finding the radius is a little more difficult. On this scale, diffraction patterns and movement towards and away from the objective complicate issues and make it more of a challenge.
Each of the droplets creates an Airy disk in the field of view, and the diameter of the central disc and the diffraction pattern can be measured. Using this knowledge, and the wavelength of the light incident on the particle, the size of the droplet can be calculated. But the issue with this is that there are two factors that can alter the size of the Airy disk: the diameter of the obscuring droplet and its distance. One of these two variables must be known in order to determine the other. In my emulsions, there are many different sizes of droplets, and a significant thickness in the flow cell, so it could be impossible to determine the radius just from measuring the Airy disk. To counter this, I am looking into the possibility of creating water droplets of a uniform and known size. With a known radius, I can then calculate the force acting on each ferritin and, if I feel bold, even track it in three dimensions.