Photoacoustic-based thermal image formation and optimization using an evolutionary genetic algorithm

• Autor(a) principal: Uliana,João Henrique
• Data de Publicação: 2018
• Outros Autores: Sampaio,Diego Ronaldo Thomaz, Carneiro,Antonio Adilton Oliveira, Pavan,Theo Zeferino
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Research on Biomedical Engineering (Online)
• Texto Completo: http://old.scielo.br/scielo.php?script=sci_arttext&pid=S2446-47402018000200147
• topic: Photoacoustic imaging; Temperature monitoring; Hyperthermia; Genetic algorithm

Abstract Introduction For improved efficiency and security in heat application during hyperthermia, it is important to monitor tissue temperature during treatments. Photoacoustic (PA) pressure wave amplitude has a temperature dependence given by the Gruenesein parameter. Consequently, changes in PA signal amplitude carry information about temperature variation in tissue. Therefore, PA has been proposed as an imaging technique to monitor temperature during hyperthermia. However, no studies have compared the performance of different algorithms to generate PA-based thermal images. Methods Here, four methods to estimate variations in PA signal amplitude for thermal image formation were investigated: peak-to-peak, integral of the modulus, autocorrelation of the maximum value, and energy of the signal. Changes in PA signal amplitude were evaluated using a 1-D window moving across the entire image. PA images were acquired for temperatures ranging from 36oC to 41oC using a phantom immersed in a temperature controlled thermal bath. Results The results demonstrated that imaging processing parameters and methods involved in tracking variations in PA signal amplitude drastically affected the sensitivity and accuracy of thermal images formation. The sensitivity fluctuated more than 20% across the different methods and parameters used. After optimizing the parameters to generate the thermal images using an evolutionary genetic algorithm (GA), the percentage of pixels within the acceptable error was improved, in average, by 7.5%. Conclusion Optimization of processing parameters using GA could increase the accuracy of measurement for this experimental setup and improve quality of PA-based thermal images.