Syracuse researchers deem noise beneficial for radiology systems
Stochastic resonance (SR), a phenomenon in which carefully selected noise is added to a system in order to enhanced sensitivity has been found to clarify signals within radar, sonar and radiography systems used in the detection of breast cancer, according to a Syracuse University (SU) research team that recently developed and patented the technology.
Hao Chen, PhD, Pramod K. Varshney, PhD, and James H. Michels, PhD, from the L.C. Smith College of Engineering and Computer Science at Syracuse, have developed a calculating method to determine the correct level and type of noise that should be added to existing noise in radiography systems.
The researchers have found that the result of the addition of noise can improve the system’s ability to identify precancerous lesions, and have noted a 36 percent reduction in the rate of false positives.
Chen explained that there is a broad spectrum of applications for the technology, noting audio, video, geophysical and environmental systems. “If a system’s performance is unsatisfactory, we add noise to the system based on a specific algorithm that can significantly improve system performance,” stated Chen.
The SR technology was recently employed in a series of mammography studies by doctoral candidate Renbin Peng, who sought to identify micro-calcifications in breast tissue. Micro-calcifications--which average approximately 0.3 mm in size and offer little contrast with surrounding tissue--are considered to be early markers for potential cancerous conditions and are typically difficult to identify. The technology allowed for improved detection of the markers, as well as a false positive rate reduced by more than one third, said the researchers.
Further research being conducted by the Syracuse group will include improvements to the efficiency of the SR-based detection techniques, said Chen and colleagues.
Hao Chen, PhD, Pramod K. Varshney, PhD, and James H. Michels, PhD, from the L.C. Smith College of Engineering and Computer Science at Syracuse, have developed a calculating method to determine the correct level and type of noise that should be added to existing noise in radiography systems.
The researchers have found that the result of the addition of noise can improve the system’s ability to identify precancerous lesions, and have noted a 36 percent reduction in the rate of false positives.
Chen explained that there is a broad spectrum of applications for the technology, noting audio, video, geophysical and environmental systems. “If a system’s performance is unsatisfactory, we add noise to the system based on a specific algorithm that can significantly improve system performance,” stated Chen.
The SR technology was recently employed in a series of mammography studies by doctoral candidate Renbin Peng, who sought to identify micro-calcifications in breast tissue. Micro-calcifications--which average approximately 0.3 mm in size and offer little contrast with surrounding tissue--are considered to be early markers for potential cancerous conditions and are typically difficult to identify. The technology allowed for improved detection of the markers, as well as a false positive rate reduced by more than one third, said the researchers.
Further research being conducted by the Syracuse group will include improvements to the efficiency of the SR-based detection techniques, said Chen and colleagues.