Colorectal cancer poses a global health risk, often heralded by colorectal polyps. Colonoscopy is the primary modality for polyp detection, with precise, real-time segmentation being key to effective diagnosis and surgical planning. Existing segmentation models like convolutional neural networks (CNNs) and Transformers have propelled progress but face trade-offs between precision and speed. CNNs excel in local feature extraction yet struggle with global context, while Transformers handle global information well but at a computational cost. Addressing these constraints, we introduce TMPSformer, a groundbreaking lightweight model tailored for efficient and accurate real-time polyp segmentation. TMPSformer, with its compact size of only 2.7 M, features a pioneering hybrid encoder merging Transformers’ long-range dependencies and shift Multi-Layer Perceptrons (MLPs)’ local dependencies, effectively enhancing segmentation performance. It also equips an All-MLP decoder to streamline feature fusion and enhance decoding efficiency. TMPSformer utilizes the Flash Efficient Attention (FEA) module to replace the traditional Attention module, significantly improving real-time performance. A comprehensive evaluation on five public polyp segmentation datasets demonstrated TMPSformer’s superiority over existing state-of-the-art algorithms. Specifically, TMPSformer achieves real-time processing at 162 frames per second (FPS) at 512 × 512 resolution on the Kvasir-SEG dataset using a single NVIDIA RTX 2080 Ti GPU, and achieves a mean Intersection over Union (mIoU) of 0.811. Its segmentation performance surpasses ColonSegNet by 8.7

The colonoscopy is the foremost technique for detecting polyps, where accurate segmentation is crucial for effective diagnosis and surgical preparation. Nevertheless, contemporary deep learning-based methods for polyp segmentation face substantial hurdles due to the large amount of labeled data required. To address this, we introduce an innovative semi-supervised learning framework based on cross-pseudo supervision (CPS) and contrastive learning, termed Semi-supervised Polyp Segmentation (SemiPolypSeg), which requires only limited labeled data. First, a new segmentation architecture, the Hybrid Transformer–CNN Segmentation Network (HTCSNet), is proposed to enhance semantic representation and segmentation performance. HTCSNet features a parallel encoder combining transformers and convolutional neural networks, as well as an All-MLP decoder with skip connections to streamline feature fusion and enhance decoding efficiency. Next, the integration of CPS in SemiPolypSeg enforces output consistency across diverse perturbed datasets and models, guided by the consistency loss principle. Finally, patch-wise contrastive loss discerns feature disparities between positive and negative sample pairs as delineated by the projector. Comprehensive evaluation demonstrated our method’s superiority over existing state-of-the-art semi-supervised segmentation algorithms. Specifically, our method achieved Dice Similarity Coefficients (DSCs) of 89.68% and 90.62% on the Kvasir-SEG dataset with 15% and 30% labeled data, respectively, and 89.72% and 90.06% on the CVC-ClinicDB dataset with equivalent ratios.

Purpose Robotic arms’ interactions with the external environment are growing more intricate, demanding higher control precision. This study aims to enhance control precision by establishing a dynamic model through the identification of the dynamic parameters of a self-designed robotic arm. Design/methodology/approach This study proposes an improved particle swarm optimization (IPSO) method for parameter identification, which comprehensively improves particle initialization diversity, dynamic adjustment of inertia weight, dynamic adjustment of local and global learning factors and global search capabilities. To reduce the number of particles and improve identification accuracy, a step-by-step dynamic parameter identification method was also proposed. Simultaneously, to fully unleash the dynamic characteristics of a robotic arm, and satisfy boundary conditions, a combination of high-order differentiable natural exponential functions and traditional Fourier series is used to develop an excitation trajectory. Finally, an arbitrary verification trajectory was planned using the IPSO to verify the accuracy of the dynamical parameter identification. Findings Experiments conducted on a self-designed robotic arm validate the proposed parameter identification method. By comparing it with IPSO1, IPSO2, IPSOd and least-square algorithms using the criteria of torque error and root mean square for each joint, the superiority of the IPSO algorithm in parameter identification becomes evident. In this case, the dynamic parameter results of each link are significantly improved. Originality/value A new parameter identification model was proposed and validated. Based on the experimental results, the stability of the identification results was improved, providing more accurate parameter identification for further applications.

The closed-form solution of each rotation joint angle is the most direct and convenient method for humanoid robot arms to imitate the real-time motion of the human arm. Occasionally, the human structure differs from that of a robot, and sensor information is inadequate. Therefore, it is difficult to obtain a closed-form solution for the inverse kinematics (IK) of a robot when using the Denavit–Hartenberg (D–H) method or screw theory alone. This study proposes combining the screw theory and the D–H method to solve the closed-form solution of the IK of a robotic arm with two unique structures: the elbow joint offset and the non-vertical state at the shoulder joint of the robotic arm. The proposed combination method first solves the elbow joint angle using screw theory. Then, the other rotation joint angles are solved using the D–H method. With the proposed method, human arm motion is successfully mapped to the corresponding joint of the robotic arm, and the closed-form solution allows the robotic arm to imitate human arm motion in real–time. Simulation results show that IK can be used to obtain high–precision angles and positions. Experiments show that the robotic arm has an accurate imitation performance, the accuracy of the $z$ -axis of the manipulator is significantly improved by 73.6%, and that of the $x$ -axis and $y$ -axis is also improved by 37.36% and 21.89%, respectively.

Coexisting ammonium (NH4+-N) and phosphate (PO43--P) in wastewater is one of the main causes of eutrophication, which poses severe risks to aquatic ecosystem and human health worldwide. Herein, magnesium-rich tricalcium aluminate (Mg/C(3)A), which was constructed by incorporating Mg into cement-based material C(3)A via solid-state reaction, was employed in the simultaneous removal of NH4+-N and PO43--P. Considering the wastewater with unbalanced N/P ratio and fluctuant pH, the effect of multiple factors (Mg/C(3)A dosage, pH, initial contaminant concentration, and temperature) on the removal of both ions were systematically investigated by employing response surface methodology technique. The results demonstrated that the impact order of the factors on the NH4+ removal by Mg/C(3)A was: temperature > Mg/C(3)A dosage > initial NH4+ concentration > pH > initial PO43- concentration; the impact order on the PO43- removal was: initial PO43- concentration > Mg/C(3)A dosage > temperature > pH > initial NH4+ concentration. The maximum removal amount of NH4+ (54.13 mg g(-1)) and PO43- (56.47 mg g(-1)) were obtained at: Mg/C(3)A dosage = 3 g L-1, initial NH4+ concentration = 160 mg L-1, initial PO43- concentration = 160 mg L-1, temperature = 308 K, and pH = 7. In addition, the possible interactive influence mechanisms were elucidated in depth. Mg2+ played a major role in the PO43- removal by forming struvite (MgNH4PO4 center dot 6H(2)O) and newberyite (MgHPO4 center dot 3H(2)O). OH released from Mg/C(3)A hydration mainly contributed to NH4+ removal. This work showed that Mg-rich C(3)A is a promising candidate for simultaneous removal of NH4+ and PO43-, shedding light on practical water remediation.

The inverse kinematics of redundant manipulators is one of the most important and complicated problems in robotics. Simultaneously, it is also the basis for motion control, trajectory planning, and dynamics analysis of redundant manipulators. Taking the minimum pose error of the end-effector as the optimization objective, a fitness function was constructed. Thus, the inverse kinematics problem of the redundant manipulator can be transformed into an equivalent optimization problem, and it can be solved using a swarm intelligence optimization algorithm. Therefore, an improved fruit fly optimization algorithm, namely, the hybrid mutation fruit fly optimization algorithm (HMFOA), was presented in this work for solving the inverse kinematics of a redundant robot manipulator. An olfactory search based on multiple mutation strategies and a visual search based on the dynamic real-time updates were adopted in HMFOA. The former has a good balance between exploration and exploitation, which can effectively solve the premature convergence problem of the fruit fly optimization algorithm (FOA). The latter makes full use of the successful search experience of each fruit fly and can improve the convergence speed of the algorithm. The feasibility and effectiveness of HMFOA were verified by using 8 benchmark functions. Finally, the HMFOA was tested on a 7-degree-of-freedom (7-DOF) manipulator. Then the results were compared with other algorithms such as FOA, LGMS-FOA, AE-LGMS-FOA, IFOA, and SFOA. The pose error of end-effector corresponding to the optimal inverse solution of HMFOA is 10−14 mm, while the pose errors obtained by FOA, LGMS-FOA, AE-LGMS-FOA, IFOA, and SFOA are 102 mm, 10−1 mm, 10−2 mm, 102 mm, and 102 mm, respectively. The experimental results show that HMFOA can be used to solve the inverse kinematics problem of redundant manipulators effectively.