Clustering-Based Particle Swarm Optimization for Electrical Impedance Imaging

Proceedings of the Genetic and Evolutionary Computation Conference

2023

A search space expansion process is proposed in the context of tomographic reconstruction (TR). The idea is to widen the effective search space in a series of increasing sizes with clamping on the search space boundary. The technique was tested on four simple phantoms and on the clinically important Shepp-Logan phantom. Dispersive flies optimisation (DFO), a lightweight particle swarm optimisation (PSO) variant, is shown to produce lower reproduction errors compared to standard TR toolbox algorithms. The expansion technique demonstrably decreases salt-and-pepper noise. DFO with 50 subspace searches was found to be superior to differential evolution, PSO and, more importantly, a number of conventional reconstruction techniques. To the best of our knowledge, this is the first work where search space expansion, in its literal form, is introduced, discussed and applied to this problem. CCS CONCEPTS• Computing methodologies → Continuous space search; • Theory of computation → Bio-inspired optimization.

Section: Introductionmentioning

confidence: 99%

Tomographic Reconstruction with Search Space Expansion

Proceedings of the Genetic and Evolutionary Computation Conference

2023

“…Batenberg [6] considered an evolutionary framework. Several authors report on swarm intelligent TR, covering areas such as: binary reconstruction [36], geophysical reconstruction [43], electrical capacitance and impedance tomography [25,47] and surface reconstruction from 3D data [16]. Additionally PART, based on the movement of particles, was proposed for binary reconstruction [4].…”

Section: Introductionmentioning

confidence: 99%

Swarm led tomographic reconstruction

Proceedings of the Genetic and Evolutionary Computation Conference

2022

Image reconstruction from ray projections is a common technique in medical imaging. In particular, the few-view scenario, in which the number of projections is very limited, is important for cases where the patient is vulnerable to potentially damaging radiation. This paper considers swarm-based reconstruction where individuals, or particles, swarm in image space in an attempt to lower the reconstruction error. We compare several swarm algorithms with standard algebraic reconstruction techniques and filtered backprojection for five standard test phantoms viewed under reduced projections. We find that although swarm algorithms do not produce solutions with lower reconstruction errors, they generally find more accurate reconstructions; that is, swarm techniques furnish reconstructions that are more similar to the original phantom. A function profiling method suggests that the ability of the swarm to optimise these high dimensional problems can be attributed to a broad funnel leading to complex structure close to the optima. This finding is further exploited by optimising the parameters of the best performing swarm technique, and the results are compared against three unconstrained and boxed local search methods. The tomographic reconstruction-optimised swarm technique is shown to be superior to prominent algebraic reconstructions and local search algorithms. CCS CONCEPTS• Computing methodologies → Continuous space search; • Theory of computation → Bio-inspired optimization.

“…Several metaheuristic algorithms have been applied to TR, including harmony search [32], tabu search [26], simulated annealing [25], memetics [16] and evolutionary algorithms [8]. Swarm algorithms have also been trialled in binary reconstruction [29], geophysical reconstruction [37], electrical capacitance and impedance tomography [38,23] and surface reconstruction from 3D data [18]. An algorithm based on the movement of particles over a single image, a pixel-swarm, has been developed for binary reconstruction [4].…”

Section: Introductionmentioning

confidence: 99%

Swarm Optimised Few-View Binary Tomography

Applications of Evolutionary Computation

2022

This paper considers a swarm optimisation approach to fewview tomographic reconstruction. DFOMAX, a high diversity swarm optimiser, demonstrably reconstructs binary images to a high fidelity, outperforming a leading algebraic technique, differential evolution and particle swarm optimisation on four standard phantoms. The paper considers the effectiveness of optimisers that have been developed for optimal low dimensional performance and concludes that trial solution clamping on the walls of the feasible search space is important for good performance.