Three-dimensional (3D) photoacoustic tomography is a method of choice for imaging round organs such as brain and breast. Many research groups have used a fully populated hemispherical transducer array with 256, 512, 1028, or 2048 elements and used that for 3D imaging. These transducer arrays are expensive and require a sophisticated data acquisition unit. Other groups have used much smaller number of transducers with a rotating mechanism which eventually filled out the entire hemisphere. We have built a 3D hemispherical array with 28 transducers which are placed on a 3D printed dome-like unit. The location of transducers however may be off-placed by a few millimeters (due to human error and errors in 3D printing). This may be to defocus the reconstructed image if the acceptable positions of transducers are not selected. In this work, we developed a compensation algorithm for misplacement of these transducers using Cuckoo search (CS) algorithm. The CS algorithm finds the optimum location for the transducers using levy flight which relies on levy distribution. The optimum location of each of these transducers is found within -4 mm to 4 mm of their locations. Universal back projection algorithm was used for image reconstruction and the sharpness of 3D image was used as the cost function; additionally, two more objective functions, the Brenner gradient, and the Tenenbaum gradient was investigated.
|