New technology for better cell observation6 March 2009 The tedium of observing cells for hours to detect changes or reactions to experiments may be a thing of past for some researchers, thanks to a new cell observation approach being developed by Monash researcher Mr Khoo Boon How. The new approach, which uses a software equipped with a learning algorithm, will be simple to use and yield better results through reduction in errors, says Khoo, a lecturer at the School of Engineering. “Understanding cell movements and behaviour is very important in research into cell embryology, wound healing, host defence mechanisms and other areas. Usually, someone has to sit and at the cells to observe their activity. This can take hours.” “Often, these changes can be difficult to detect because they’re happening in a very dynamic environment, thus leading to errors. Therefore, a system like this is very useful in capturing accurate results from experiments.” Khoo says there are several solutions currently available in the market, but few are able to monitor more than one type of cell at one time. He says the system being developed by Khoo is robust and can deal with changes in shape, cell content, environment and position. Researchers can also track any type of cell in real time, with results available immediately on completion of the experiment. Khoo says a key feature of the monitoring system is its ability to undergo supervised learning - where the software is basically 'taught' what to observe and track. “The same learning process can be used for multiple cells. If learning is good, then it will be a robust system and we don’t have to train it for observation of similar cells in the future.” Khoo says the current testing of the system indicates that is robust enough to track cells regardless of a change in orientation, cell content and shape. The system has also successfully tracked multiple cells with good accuracy. “The best thing about the system is its simplicity – all one needs is a microscope, camera, a firewire card and computer.” Work now, he says, is focused on enhancing the system’s learning capabilities, reducing the processing time and adding a very useful additional feature – removing irrelevant imagery from the end result. - Jasbir SinghFor more information:
|
. |
The diagram above outlines the process involved in breaking down a cell image for processing and identification.
The system is able to track multiple types of cells, regardless of their change in orientation, cell content, and shape.
|