Simultaneous localization of multiple tumors from thermogram of tissue phantom by using a novel optimization algorithm inspired by hunting dogs.

The objective of this study is to couple the contact thermography method with a novel optimization algorithm to rapidly detect and localize the soft tissue tumor. To this end, experiments are carried out on tissue-mimicking phantoms containing resistance heaters to simulate the embedded tumors. An examiner robot is used to measure the temperature of the tissue surface. The time required for the examination of the tissue surface is reduced by developing a novel optimization algorithm called the Hunter Algorithm (HA). In the HA, population individuals are called the hunters, and the global maximum is referred to as the prey. The maximum temperature occurs at the location of the tumor. By the end of the hunting procedure, a flock of hunters converges to the maximum temperature and reaches the tumor while the examination time is significantly reduced. Performance of the HA is evaluated by applying the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO) algorithm to 11 test functions as minimization problems. It is observed that for the Ackley's function, as an example, the HA finds the global minimum after the 10th iteration with an accuracy of 10-4, while the PSO converges with the same accuracy after the 30th iteration and the accuracy of the GA remains about 0.002. In addition, the results show that the contact thermography in conjunction with the HA is of clinical importance in accurate detection of multiple tumors and small and deeply located tumors with insignificant thermal effects on the tissue surface.