Artificial Immunealgorithm And Itsapplications In Nuclear Molecular Imaging

Artificial immune algorithm is a bio-inspired soft computing that uses concepts and ideas from the immune system theory to solve problems of engineering and scientific applications. It has been a characteristic research field based on the principles of artificial immune system. This paper focuses on the immune clonal selection algorithm and artificial immune network algorithm for optimaization in quantitative analysis of nuclear medicine molecular imaging. This paper has the following main aspects:First, this paper studied the PKAIN (artificial immune network for parameter optimization of pharmacokinetics) algorithm and extended it to solve parameter optimization of tracer kinetic models for molecular imaging. In dynamic imaging experiments of 18F-FDG metabolism with small animal PET, tracer time activity curves (TACs) in several regions were obtained. PKAIN was adapted here to optimize parameters of the tracer kinetic models. The experiment results showed that the PKAIN algorithm had the ability to get better solution than kinetic imaging system (KIS) and particle swarm optimization (PSO) algorithm. It gave a novel view to obtain more accurate and reliable kinetic models of molecular imaging.Second, the improved PKAIN was used to optimize parameters of tracer kinetic models. Based on cooperative theory, an improved PKAIN artificial immune network with cooperative operator was developed for optimization. Due to the global swarm cooperative operator, memory cells with particle swarm behavior are capable of sharing search experience. Furthermore, one dimension mutation (odm) operater was proposed. Experimental results indicated that the cooperative operater improve searching ability.Third, simultaneous estimation (SIME) of tracer kinetic parameters weighted multi-objective problem into a single target traditionally. Single-objective optimization algorithm was used to obtain parameters of the model traditionally. In this paper, a new application-driven algorithm called PKICA (immune clonal algorithm for parameter optimization of pharmacokinetics) was proposed to solve multi-objective problems. And PKICA was used in SIME. A series of operators were designed followed clonal selection theory. Classical test functions were used here to assess the performance of the PKICA. Compared with PSO algorithm and genetic algorithm, the PKICA presented good performance with desirable global searching ability, convergence property and diversity property. Multi-objective optimization approach outperformed traditional single-objective approach according to SIME results.