A study by researchers at Syracuse University’s L.C. Smith College of Engineering and Computer Science demonstrated that the addition of noise, something that is usually an extra annoyance to mammography and radiography systems, enhances system precision.

The study explores a phenomenon called stochastic resonance, which makes cancerous lesions more detectable by adding noise to the mammogram system, said Hao Chen, an associate professor of electrical and computer engineering.

‘For any given mammograph, there are normal tissues and lesion tissues,’ Chen said. ‘An algorithm locates and identifies the lesion tissues. By adding noise to existing algorithms, we have a more accurate location of the lesions.’

The study was led by Chen; Pramod K. Varshney, a professor of electrical and computer engineering; and James Michels, a research professor. In 2006, Renbin Peng, a research assistant, joined them in their work on the application of this technology to mammography, and they received the patent for their research in July 2009.

The sound added to the system is actually a random signal for detecting lesions, rather than what we traditionally think of as sound, Varshney said.

The researchers have worked to carefully select the most effective noise to improve the detection of cancerous lesions. Their research works on which noise achieves the best possible mammogram performance, Chen said.

The increase in precision of the mammograph will be of particular importance to physicians because they are likely form a more accurate diagnosis, Chen said.

‘If you have a surgery, and you thought it was a lesion, and it turns out to be normal tissue, that’s devastating,’ Chen said. ‘Our algorithm reduces this kind of probability because lesion detection is better. We are able to identify more lesions at the cost of less misdiagnosis.’

Adding this sound technology has proven to be particularly helpful in reducing false positives in a mammograph image and improving detection of areas that contain an anomaly, Varshney said.

The researchers hope to further develop their research of this phenomenon so that they can combine their engineering expertise with the knowledge of physicians and researchers in the medical field to produce more useful applications, Chen said.

‘I think the next step is for some industries to get involved, adopt the technology and get to the prototyping stage. This will improve the quality of image so that the machine or humans can (improve diagnoses),’ Varshney said.

Systems such as hearing aids like cochlear implants and hearing devices use this kind of spontaneous noise, which improves an individual’s hearing, Chen said.

But not all systems will improve with this technology, Varshney said. Radios and electronic devices such as the iPod wouldn’t work better with this signal technology, he said.

‘Hopefully someday we will be able to make a real impact. This technology has a very bright future because it’s adaptive, low cost and it can improve the systems which are not replaceable,’ he said.

vdnapoli@syr.edu

Published on January 31, 2010 at 12:00 pm