Glioma is the most common type of brain tumor, and it has a high mortality rate. Accurate tumor segmentation based on magnetic resonance imaging (MRI) is of great significance for the diagnosis and treatment of brain tumors. Recently, a brain tumor segmentation method based on U-Net has gained considerable attention. However, brain tumor segmentation is a challenging task due to the structural variations and inhomogeneous intensity of tumors. Existing brain tumor segmentation studies have shown that the problems of insufficient down-sampling feature extraction and loss of up-sampling information arise when using U-Net to segment brain tumors. In this study, we proposed an improved U-Net model, SEResU-Net, which combines the deep residual network and the Squeeze-and-Excitation Network (SENet). The deep residual network solves the problem of network degradation so that SEResU-Net can extract more feature information. SENet avoids information loss and enables the network to focus on the useful feature map, which solves the problem of insufficient segmentation accuracy of small-scale brain tumors. Furthermore, we developed a fusion loss function, which is composed of Dice loss and cross-entropy loss, to solve the problems of network convergence and data imbalance. We evaluated the segmentation performance of SEResU-Net on BraTS2018 and BraTS2019, two authoritative MRI brain tumor benchmarks. Experimental results revealed that SEResU-Net outperformed SEU-Net, ResU-Net, and U-Net. For SEResU-Net, the mean Dice similarity coefficients were 0.9373, 0.9108, and 0.8758 for the whole tumor, the tumor core, and the enhanced tumor, which were 7.10%, 11.88%, and 15.33% greater than those of the U-Net benchmark network, respectively. Our findings demonstrate that the proposed SEResU-Net has a competitive effect in segmenting multimodal brain tumors.

We propose a new computing-inspired bio-detection framework called touchable computing (TouchComp). Under the rubric of TouchComp, the best solution is the cancer to be detected, the parameter space is the tissue region at high risk of malignancy, and the agents are the nanorobots loaded with contrast medium molecules for tracking purpose. Subsequently, the cancer detection procedure (CDP) can be interpreted from the computational optimization perspective: a population of externally steerable agents (i.e., nanorobots) locate the optimal solution (i.e., cancer) by moving through the parameter space (i.e., tissue under screening), whose landscape (i.e., a prescribed feature of tissue environment) may be altered by these agents but the location of the best solution remains unchanged. One can then infer the landscape by observing the movement of agents by applying the “seeing-is-sensing” principle. The term “touchable” emphasizes the framework’s similarity to controlling by touching the screen with a finger, where the external field for controlling and tracking acts as the finger. Given this analogy, we aim to answer the following profound question: can we look to the fertile field of computational optimization algorithms for solutions to achieve effective cancer detection that are fast, accurate, and robust? Along this line of thought, we consider the classical particle swarm optimization (PSO) as an example and propose the PSO-inspired CDP, which differs from the standard PSO by taking into account realistic in vivo propagation and controlling of nanorobots. Finally, we present comprehensive numerical examples to demonstrate the effectiveness of the PSO-inspired CDP for different blood flow velocity profiles caused by tumor-induced angiogenesis. The proposed TouchComp bio-detection framework may be regarded as one form of natural computing that employs natural materials to compute.

We propose a new framework of computing-inspired multifocal cancer detection procedure (MCDP). Under the rubric of MCDP, the tumor foci to be detected are regarded as solutions of the objective function, the tissue region around the cancer areas represents the parameter space, and the nanorobots loaded with contrast medium molecules for cancer detection correspond to the optimization agents. The process that the nanorobots detect tumors by swimming in the high-risk tissue region can be regarded as the process that the agents search for the solutions of an objective function in the parameter space with some constraints. For multimodal optimization (MMO) aiming to locate multiple optimal solutions in a single simulation run, the niche technology has been widely used. Specifically, the niche genetic algorithm (NGA) has been shown to be particularly effective in solving MMO. It can be used to identify the global optima of multiple hump functions in a running, effectively keep the diversity of the population, and prematurely avoid the genetic algorithm. Learning from the optimization procedure of NGA, we propose the NGA-inspired MCDP in order to locate the tumor targets efficiently while taking into account realistic in vivo propagation and controlling of nanorobots, which is different from the use scenario of the standard NGA. To improve the performance of the MCDP, we also modify the crossover operator of the original NGA from crossing within a population to crossing between two populations. Finally, we present comprehensive numerical examples to demonstrate the effectiveness of the NGA-inspired MCDP when the biological objective function is associated with the blood flow velocity profile caused by tumor-induced angiogenesis.

With the development of nanotechnology, externally manipulable or self-regulatable smart nanosystems can be utilized as effective tools for computational nanobiosensing, where natural computing strategies are exploited for knowledge-aided nanobiosensing.

Several evolution strategies for in vivo computation are proposed with the aim of realizing tumor sensitization and targeting (TST) by externally manipulable nanoswimmers. In such targeting systems, nanoswimmers assembled by magnetic nanoparticles are externally manipulated to search for the tumor in the high-risk tissue by a rotating magnetic field produced by a coil system. This process can be interpreted as in vivo computation, where the tumor in the high-risk tissue corresponds to the global maximum or minimum of the in vivo optimization problem, the nanoswimmers are seen as the computational agents, the tumor-triggered biological gradient field (BGF) is used for fitness evaluation of the agents, and the high-risk tissue is the search space. Considering that the state-of-the-art magnetic nanoswimmer control method can only actuate all the nanoswimmers heading in the same direction simultaneously, we introduce the orthokinetic movement strategies into the agent location updating in the existing swarm intelligence algorithms. Especially, the gravitational search algorithm (GSA) is revisited and the corresponding in vivo optimization algorithm called orthokinetic GSA (OGSA) is proposed to carry out the TST. Furthermore, to determine the direction of the orthokinetic agent movement in every iteration of the operation, we propose several strategies according to the fitness ranking of the nanoswimmers in the BGF. To verify the superiority of the OGSA and choose the optimal evolution strategy, some numerical experiments are presented and compared with that of the brute-force search, which represents the traditional method for TST. It is found that the TST performance can be improved by the weak priority evolution strategy (WP-ES) in most of the scenarios.