CCDet: Confidence-Consistent Learning for Dense Object Detection.

Modern detectors commonly employ classification scores to reflect the localization quality of detection results. However, there exists an inconsistency between them, misguiding the selection of high-quality predictions and providing unreliable results for downstream applications. In this paper, we find that the root of this confidence inconsistency lies in the inaccurate IoU estimation and the spatial misalignment of the learned features between the classification and localization tasks. Therefore, a Confidence-Consistent Detector (CCDet) which includes the Distribution-based IoU Prediction (DIP) and Consistency-aware label assignment (CLA), is proposed. DIP provides more stable and accurate IoU estimation by learning the probability distribution over the IoU range and employing the expectation as the predicted IoU. CLA adopts both the prediction performance and consistency degree of samples as assignment metrics to select positives, which guides the classification and localization tasks to promote similar feature distribution. Comprehensive experiments demonstrate that CCDet can effectively mitigate the confidence inconsistency between classification and localization, and achieve stable improvement across different baselines. On the test-dev set of MS COCO, CCDet acquires a single-model single-scale AP of 50.1%, surpassing most of the existing object detectors.

A magnetic multi-layer soft robot for on-demand targeted adhesion.

Magnetic soft robots have shown great potential for biomedical applications due to their high shape reconfigurability, motion agility, and multi-functionality in physiological environments. Magnetic soft robots with multi-layer structures can enhance the loading capacity and function complexity for targeted delivery. However, the interactions between soft entities have yet to be fully investigated, and thus the assembly of magnetic soft robots with on-demand motion modes from multiple film-like layers is still challenging. Herein, we model and tailor the magnetic interaction between soft film-like layers with distinct in-plane structures, and then realize multi-layer soft robots that are capable of performing agile motions and targeted adhesion. Each layer of the robot consists of a soft magnetic substrate and an adhesive film. The mechanical properties and adhesion performance of the adhesive films are systematically characterized. The robot is capable of performing two locomotion modes, i.e., translational motion and tumbling motion, and also the on-demand separation with one side layer adhered to tissues. Simulation results are presented, which have a good qualitative agreement with the experimental results. The feasibility of using the robot to perform multi-target adhesion in a stomach is validated in both ex-vivo and in-vivo experiments.

An Active Vibration Isolation and Compensation System for Improving Optical Image Quality: Modeling and Experiment.

Optical detection equipment (ODE) is subjected to vibrations that hamper the quality of imaging. In this paper, an active vibration isolation and compensation system (VICS) for the ODE is developed and systematically studied to improve the optical imaging quality. An active vibration isolator for cameras is designed, employing a dual-loop control strategy with position compensation and integral force feedback (IFF) control, and establishing the mapping relationship between vibration and image quality. A performance metric for evaluating images is also proposed. Finally, an experimental platform is constructed to verify its effectiveness. Based on the experimental results, it can be concluded that the proposed VICS effectively isolates vibrations, resulting in a reduction of 13.95 dB in the peak at the natural frequency and an 11.76 Hz widening of the isolation bandwidth compared with the system without it. At the same time, the experiments demonstrate that the image performance metric value increases by 46.03% near the natural frequency.

Structure-Crack Detection and Digital Twin Demonstration Based on Triboelectric Nanogenerator for Intelligent Maintenance.

The accomplishment of condition monitoring and intelligent maintenance for cantilever structure-based energy harvesting devices remains a challenge. Here, to tackle the problems, a novel cantilever-structure freestanding triboelectric nanogenerator (CSF-TENG) is proposed, which can capture ambient energy or transmit sensory information. First, with and without a crack in cantilevers, the simulations are carried out. According to simulation results, the maximum change ratios of natural frequency and amplitude are 1.1% and 2.2%, causing difficulties in identifying defects by these variations. Thus, based on Gramian angular field and convolutional neural network, a defect detection model is established to achieve the condition monitoring of the CSF-TENG, and the experimental result manifests that the accuracy of the model is 99.2%. Besides, the relation between the deflection of cantilevers and the output voltages of the CSF-TENG is first built, and then the defect identification digital twin system is successfully created. Consequently, the system is capable of duplicating the operation of the CSF-TENG in a real environment, and displaying defect recognition results, so the intelligent maintenance of the CSF-TENG can be realized.

Electrochromic-Induced Rechargeable Aqueous Batteries: An Integrated Multifunctional System for Cross-Domain Applications.

Multifunctional electrochromic-induced rechargeable aqueous batteries (MERABs) integrate electrochromism and aqueous ion batteries into one platform, which is able to deliver the conversion and storage of photo-thermal-electrochemical sources. Aqueous ion batteries compensate for the drawbacks of slow kinetic reactions and unsatisfied storage capacities of electrochromic devices. On the other hand, electrochromic technology can enable dynamically regulation of solar light and heat radiation. However, MERABs still face several technical issues, including a trade-off between electrochromic and electrochemical performance, low conversion efficiency and poor service life. In this connection, novel device configuration and electrode materials, and an optimized compatibility need to be considered for multidisciplinary applications. In this review, the unique advantages, key challenges and advanced applications are elucidated in a timely and comprehensive manner. Firstly, the prerequisites for effective integration of the working mechanism and device configuration, as well as the choice of electrode materials are examined. Secondly, the latest advances in the applications of MERABs are discussed, including wearable, self-powered, integrated systems and multisystem conversion. Finally, perspectives on the current challenges and future development are outlined, highlighting the giant leap required from laboratory prototypes to large-scale production and eventual commercialization.

A cantilever-structure triboelectric nanogenerator for energy harvesting and defect detection via long short-term memory network.

The electric signals of cantilever energy harvesting devices with/without a crack were mainly obtained by external sensors, so detecting device damage on a large scale is difficult. To tackle the issue, a cantilever-structure freestanding triboelectric nanogenerator (CSF-TENG) device was proposed, which can scavenge ambient energy and act as a self-powered sensor. Firstly, the relation between the peak-to-peak voltage and amplitude of the CSF-TENG was established. Next, the output performance of the CSF-TENG was measured. Then, depending on electric signals output by the CSF-TENG, a cantilever defect identification model was built by using the wavelet packet and long short-term memory (LSTM) algorithms. The experimental results manifest that the accuracy of the model is about 98.6%. Thus, the CSF-TENG with a crack can be detected timely due to its self-monitoring ability, which is of great significance for the development of self-powered sensor networks.

Power Density Improvement of Piezoelectric Energy Harvesters via a Novel Hybridization Scheme with Electromagnetic Transduction.

A novel hybridization scheme is proposed with electromagnetic transduction to improve the power density of piezoelectric energy harvester (PEH) in this paper. Based on the basic cantilever piezoelectric energy harvester (BC-PEH) composed of a mass block, a piezoelectric patch, and a cantilever beam, we replaced the mass block by a magnet array and added a coil array to form the hybrid energy harvester. To enhance the output power of the electromagnetic energy harvester (EMEH), we utilized an alternating magnet array. Then, to compare the power density of the hybrid harvester and BC-PEH, the experiments of output power were conducted. According to the experimental results, the power densities of the hybrid harvester and BC-PEH are, respectively, 3.53 mW/cm3 and 5.14 µW/cm3 under the conditions of 18.6 Hz and 0.3 g. Therefore, the power density of the hybrid harvester is 686 times as high as that of the BC-PEH, which verified the power density improvement of PEH via a hybridization scheme with EMEH. Additionally, the hybrid harvester exhibits better performance for charging capacitors, such as charging a 2.2 mF capacitor to 8 V within 17 s. It is of great significance to further develop self-powered devices.

Fuzzy Control and Filtering for Nonlinear Singularly Perturbed Markov Jump Systems.

This article addresses the H∞ control and filtering problems for Markov jump singularly perturbed systems approximated by Takagi-Sugeno fuzzy models. The underlying transition probabilities (TPs) are assumed to vary randomly in a finite set, which is characterized by a higher level TP matrix. The mode- and variation-dependent fuzzy static output-feedback controller (SOFC) and filter are designed, respectively, to fulfill the control and filtering purposes. To facilitate the fuzzy SOFC synthesis, the closed-loop system is transformed into a fuzzy piecewise-homogeneous Markov jump singularly perturbed descriptor system (MJSPDS) by descriptor representation. A rigorous proof of mean-square exponential admissibility for the resulting fuzzy MJSPDS is presented. The criterion ensuring the mean-square exponential stability of the fuzzy filtering error system is further formed based on similar procedures. By setting the specific forms of the related matrix variables, the solutions for the predesigned fuzzy SOFC and filter are furnished, respectively. Finally, feasibility and validities of the developed fuzzy control and filtering results are verified by two practical examples.

Piezoelectric Energy Harvesting from Suspension Structures with Piezoelectric Layers.

In this paper, we propose a generator for piezoelectric energy harvesting from suspension structures. This device consists of a leaf spring and eight pairs of piezoelectric layers attached to inner and outer surfaces. We present a special type of leaf spring, which can magnify the force from the workload to allow the piezoelectric layers to achieve larger deformation. The generator is to solve the problem of vibration energy reutilization in a low-frequency vibration system. To verify the efficiency of the proposed configuration, a series of experiments are operated. The results indicate that the resonance frequency (25.2 Hz) obtained from the sweep experiment is close to the simulation result (26.1 Hz). Impedance-matching experiments show that the sum of the output power attains 1.7 mW, and the maximum single layer reaches 0.6 mW with an impedance matching of 610 KΩ, and the instantaneous peak-peak power density is 3.82 mW/cm3. The capacitor-charging performance of the generator is also excellent under the series condition. For a 4.7 µF capacitor, the voltage is charged to 25 V in 30 s and limited at 32 V in 80 s. These results demonstrate the exploitable potential of piezoelectric energy harvesting from suspension structures.

Large-Scale Place Recognition Based on Camera-LiDAR Fused Descriptor.

In the field of autonomous driving, carriers are equipped with a variety of sensors, including cameras and LiDARs. However, the camera suffers from problems of illumination and occlusion, and the LiDAR encounters motion distortion, degenerate environment and limited ranging distance. Therefore, fusing the information from these two sensors deserves to be explored. In this paper, we propose a fusion network which robustly captures both the image and point cloud descriptors to solve the place recognition problem. Our contribution can be summarized as: (1) applying the trimmed strategy in the point cloud global feature aggregation to improve the recognition performance, (2) building a compact fusion framework which captures both the robust representation of the image and 3D point cloud, and (3) learning a proper metric to describe the similarity of our fused global feature. The experiments on KITTI and KAIST datasets show that the proposed fused descriptor is more robust and discriminative than the single sensor descriptor.

Simultaneously Quantifying Both Young's Modulus and Specific Membrane Capacitance of Bladder Cancer Cells with Different Metastatic Potential.

Both Young's modulus and specific membrane capacitance (SMC) are two important physical parameters for characterizing cell status. In this paper, we utilized a thin-neck-micropipette aspiration system to simultaneously quantify Young's modulus and SMC value of six types of cell lines in different progression grades, which include four grades from the lowest metastatic potential G1 to the highest potential G4. We investigated how these two physical properties possess heterogeneities in bladder cancer cells with different grades and what roles they might play in grading bladder cancer. The characterization results of these cells of different cancer grades is linearly correlated with the cancer grades, showing that the Young's modulus is negatively linearly correlated with bladder cancer grades, while SMC shows a positive linear correlation. Furthermore, the combination of these two physical properties on a scatter diagram clearly shows the cell groups with different cancer grades, which means that this combination could be a potential tumor grading marker to identify cancer cells with different metastatic potential.

On-Chip Construction of Multilayered Hydrogel Microtubes for Engineered Vascular-Like Microstructures.

Multilayered and multicellular structures are indispensable for constructing functional artificial tissues. Engineered vascular-like microstructures with multiple layers are promising structures to be functionalized as artificial blood vessels. In this paper, we present an efficient method to construct multilayer microtubes embedding different microstructures based on direct fabrication and assembly inside a microfluidic device. This four-layer microfluidic device has two separate inlets for fabricating various microstructures. We have achieved alternating-layered microtubes by controlling the fabrication, flow, and assembly time of each microstructure, and as well, double-layered microtubes have been built by a two-step assembly method. Modifications of both the inner and outer layers was successfully demonstrated, and the flow conditions during the on-chip assembly were evaluated and optimized. Each microtube was successfully constructed within several minutes, showing the potential applications of the presented method for building engineered vascular-like microstructures with high efficiency.

A Novel Biodegradable Multilayered Bioengineered Vascular Construct with a Curved Structure and Multi-Branches.

Constructing tissue engineered vascular grafts (TEVG) is of great significance for cardiovascular research. However, most of the fabrication techniques are unable to construct TEVG with a bifurcated and curved structure. This paper presents multilayered biodegradable TEVGs with a curved structure and multi-branches. The technique combined 3D printed molds and casting hydrogel and sacrificial material to create vessel-mimicking constructs with customizable structural parameters. Compared with other fabrication methods, the proposed technique can create more native-like 3D geometries. The diameter and wall thickness of the fabricated constructs can be independently controlled, providing a feasible approach for TEVG construction. Enzymatically-crosslinked gelatin was used as the material of the constructs. The mechanical properties and thermostability of the constructs were evaluated. Fluid-structure interaction simulations were conducted to examine the displacement of the construct's wall when blood flows through it. Human umbilical vein endothelial cells (HUVECs) were seeded on the inner channel of the constructs and cultured for 72 h. The cell morphology was assessed. The results showed that the proposed technique had good application potentials, and will hopefully provide a novel technological approach for constructing integrated vasculature for tissue engineering.

Micropipette Aspiration of Single Cells for Both Mechanical and Electrical Characterization.

Cellular physical properties have been identified to reflect cell states. Existing techniques are able to characterize either mechanical or electrical properties of a cell. This paper presents a micropipette aspiration technique that enables the characterization of both mechanical (instantaneous elastic modulus, equilibrium elastic modulus, and viscosity), and electrical (specific membrane capacitance) properties of the same single cell. Two bladder cancer cell lines (RT4 and T24) with different metastatic potential were used to evaluate the technique. The results showed that high-grade bladder cancer cells (T24, grade III) possess lower viscosity, lower elastic modulus, and larger SMC than the low-grade cancer cells (RT4, grade I). The Naive Bayes classifier was utilized to assess the classification accuracy using single-physical and multi-physical parameters. The classification results confirmed that the use of multi-biophysical parameters resulted in higher accuracy (97.5%), sensitivity (100%), and specificity (95.2%) than the use of a single-physical parameter for distinguishing T24 and RT4 cells.

Construction of All-in-Focus Images Assisted by Depth Sensing.

Multi-focus image fusion is a technique for obtaining an all-in-focus image in which all objects are in focus to extend the limited depth of field (DoF) of an imaging system. Different from traditional RGB-based methods, this paper presents a new multi-focus image fusion method assisted by depth sensing. In this work, a depth sensor is used together with a colour camera to capture images of a scene. A graph-based segmentation algorithm is used to segment the depth map from the depth sensor, and the segmented regions are used to guide a focus algorithm to locate in-focus image blocks from among multi-focus source images to construct the reference all-in-focus image. Five test scenes and six evaluation metrics were used to compare the proposed method and representative state-of-the-art algorithms. Experimental results quantitatively demonstrate that this method outperforms existing methods in both speed and quality (in terms of comprehensive fusion metrics). The generated images can potentially be used as reference all-in-focus images.

A Novel Method for Extrinsic Calibration of Multiple RGB-D Cameras Using Descriptor-Based Patterns.

This paper presents a novel method to estimate the relative poses between RGB-D cameras with minimal overlapping fields of view. This calibration problem is relevant to applications such as indoor 3D mapping and robot navigation that can benefit from a wider field of view using multiple RGB-D cameras. The proposed approach relies on descriptor-based patterns to provide well-matched 2D keypoints in the case of a minimal overlapping field of view between cameras. Integrating the matched 2D keypoints with corresponding depth values, a set of 3D matched keypoints are constructed to calibrate multiple RGB-D cameras. Experiments validated the accuracy and efficiency of the proposed calibration approach.

Robotic Immobilization of Motile Sperm for Clinical Intracytoplasmic Sperm Injection.

OBJECTIVE: In clinical intracytoplasmic sperm injection (ICSI), a motile sperm must be immobilized before insertion into an oocyte. This paper aims to develop a robotic system for automated tracking, orientation control, and immobilization of motile sperms for clinical ICSI applications. METHODS: We adapt the probabilistic data association filter by adding sperm head orientation into state variables for robustly tracking the sperm head and estimating sperm tail positions under interfering conditions. The robotic system also utilizes a motorized rotational microscopy stage and a new visual servo control strategy that predicts and compensates for sperm movements to actively adjust sperm orientation for immobilizing a sperm swimming in any direction. RESULTS: The system robustly tracked sperm head with a tracking success rate of 96.0% and estimated sperm tail position with an accuracy of 1.08 µm under clinical conditions where the occlusion of the target sperm and interference from other sperms occur. Experimental results from robotic immobilization of 400 sperms confirmed that the system achieved a consistent immobilization success rate of 94.5%, independent of sperm velocity or swimming direction. CONCLUSION: Our adapted tracking algorithm effectively distinguishes the target sperm from interfering sperms. Predicting and compensating for sperm movements significantly reduce the positioning error during sperm orientation control. These features make the robotic system suitable for automated sperm immobilization. SIGNIFICANCE: The robotic system eliminates stringent skill requirements in manual sperm immobilization. It is capable of manipulating sperms swimming in an arbitrary direction with a high success rate.

Effect of Cell Inner Pressure on Deposition Volume in Microinjection.

Microinjection is a widely used technique for introducing exogenous materials into cells. Many applications of microinjection, such as gene editing and drug testing, rely on the accurate control of the deposition volume. However, the deposition volume in microinjection is presently calibrated in an open medium without considering the cell inner pressure effect, which we experimentally show in this paper that it can induce an error as large as 30% between the actual deposition volume and the set volume. In this work, the relationship between the cell inner pressure and the deposition volume was analytically modeled and experimentally validated. On the basis of the developed model, the cell inner pressure of a given cell type can be well estimated from the injection pressure and the resulting deposition volume. The quantitated cell inner pressure is then used to reduce the error between the set volume and the actual deposition volume. Experiments conducted on human bladder cancer cells (T24 and RT4) showed that T24 cells have a higher inner pressure than RT4 cells (405 ± 45 Pa vs 341 ± 34 Pa), and after compensating for the cell inner pressure, the error between the intended set volume and the actual deposition volume into a cell became less than 3%.

Automated Non-Invasive Measurement of Single Sperm's Motility and Morphology.

Measuring cell motility and morphology is important for revealing their functional characteristics. This paper presents automation techniques that enable automated, non-invasive measurement of motility and morphology parameters of single sperm. Compared to the status quo of qualitative estimation of single sperm's motility and morphology manually, the automation techniques provide quantitative data for embryologists to select a single sperm for intracytoplasmic sperm injection. An adapted joint probabilistic data association filter was used for multi-sperm tracking and tackled challenges of identifying sperms that intersect or have small spatial distances. Since the standard differential interference contrast (DIC) imaging method has side illumination effect which causes inherent inhomogeneous image intensity and poses difficulties for accurate sperm morphology measurement, we integrated total variation norm into the quadratic cost function method, which together effectively removed inhomogeneous image intensity and retained sperm's subcellular structures after DIC image reconstruction. In order to relocate the same sperm of interest identified under low magnification after switching to high magnification, coordinate transformation was conducted to handle the changes in the field of view caused by magnification switch. The sperm's position after magnification switch was accurately predicted by accounting for the sperm's swimming motion during magnification switch. Experimental results demonstrated an accuracy of 95.6% in sperm motility measurement and an error <10% in morphology measurement.

A Flexure-Guided Piezo Drill for Penetrating the Zona Pellucida of Mammalian Oocytes.

Mammalian oocytes such as mouse oocytes have a highly elastic outer membrane, zona pellucida (ZP) that cannot be penetrated without significantly deforming the oocyte, even with a sharp micropipette. Piezo drill devices leverage lateral and axial vibration of the micropipette to accomplish ZP penetration with greatly reduced oocyte deformation. However, existing piezo drills all rely on a large lateral micropipette vibration amplitude ( 20 ) and a small axial vibration amplitude (0.1 ). The very large lateral vibration amplitude has been deemed to be necessary for ZP penetration although it also induces larger oocyte deformation and more oocyte damage. This paper reports on a new piezo drill device that uses a flexure guidance mechanism and a systematically designed pulse train with an appropriate base frequency. Both simulation and experimental results demonstrate that a small lateral vibration amplitude (e.g., 2 ) and an axial vibration amplitude as large as 1.2 were achieved. Besides achieving 100% effectiveness in the penetration of mouse oocytes (n = 45), the new piezo device during ZP penetration induced a small oocyte deformation of 3.4 versus larger than 10 using existing piezo drill devices.