Mingyu Wang

Bio: Mingyu Wang is an academic researcher from Soochow University (Suzhou). The author has contributed to research in topics: Scanning electron microscope & Calibration (statistics). The author has an hindex of 1, co-authored 8 publications receiving 5 citations.

Visual Servoing-Based Nanorobotic System for Automated Electrical Characterization of Nanotubes inside SEM.

Abstract: The maneuvering and electrical characterization of nanotubes inside a scanning electron microscope (SEM) has historically been time-consuming and laborious for operators. Before the development of automated nanomanipulation-enabled techniques for the performance of pick-and-place and characterization of nanoobjects, these functions were still incomplete and largely operated manually. In this paper, a dual-probe nanomanipulation system vision-based feedback was demonstrated to automatically perform 3D nanomanipulation tasks, to investigate the electrical characterization of nanotubes. The XY-position of Atomic Force Microscope (AFM) cantilevers and individual carbon nanotubes (CNTs) were precisely recognized via a series of image processing operations. A coarse-to-fine positioning strategy in the Z-direction was applied through the combination of the sharpness-based depth estimation method and the contact-detection method. The use of nanorobotic magnification-regulated speed aided in improving working efficiency and reliability. Additionally, we proposed automated alignment of manipulator axes by visual tracking the movement trajectory of the end effector. The experimental results indicate the system’s capability for automated measurement electrical characterization of CNTs. Furthermore, the automated nanomanipulation system has the potential to be extended to other nanomanipulation tasks.

Cooperation Method of Symmetrically Distributed Multi-Nanorobotic Manipulators Inside SEM for Nanodevice Constructing

Abstract: This paper presented a series of cooperation strategies of symmetrically distributed multi-nanorobotic manipulators according to different tasks during nanodevice fabrication inside scanning electron microscopy (SEM). For constructing carbon nanotube (CNT) based nanodevice such as carbon nanotube field effect transistor (CNTFET), the mainly assembly processes implemented by different nanorobotic manipulators were discussed. Manipulation strategies for each task during assembly were established based on nanorobotic manipulation system. Experiments were designed and carried out to evaluate the effectiveness of different manipulation strategies designed for multi-task. The results shown that the designed manipulation strategies was adapted for the nanorobotic manipulation system during CNT based nanodevice assembly.

System Calibration Towards Automated Nanomanipulation Inside Scanning Electron Microscope

Abstract: This paper presented the calibration of the nanorobotic manipulator inside Scanning Electron Microscope (SEM). Repeating movement is frequently happing to reach the predetermined position and in order to realize the automated nanomanipulation inside SEM, the errors of the nanomanipultor and the system must be analyzed. A series of experiments based on SEM image feedback were conducted and the end-effectors can characterize the relationship of the two axis. An atomic force microscope (AFM) cantilever was assembled on the top of the nanorobotic manipulator, which moved straightly along the X or Y axis under stable speed. Through SEM video stream, a image processing method was used to analyze the errors of the nanomanipultor and the thermal drift of the encoder. The movement of nanomanipultor presented the good linearity and the error between the platform and the image coordinates were deteced. The slope of the x-direction is less than 1.7° and the slope of the y-direction is less than 0.8°. The nanomanipultor corrected his own track until the drift error is accumulated to 260 nm.

Scanning Electron Microscope Calibration with SE2 and Inlens Detectors

Abstract: In this paper, we proposed an ideal that calibration in the scanning electron microscope (SEM) with SE2 detector and Inlens detector. The parameters (intrinsic and extrinsic) were both achieved. This approach required rotation and translation of the chessboard calibration to obtain multi-images. Experiments were realized by varying the orientation and the position of chessboard pattern from different work distance (WD). It can be seen from the calibration results that the SE2 detector and the Inlens detector have different overall average pixels at different work distances. By comparing the calibration results, it was found that the two detectors had close pixel errors when the work distance was between 6.4mm and 6.5mm. The results show that the calibration approach was accurate and efficient.

Adaptive Threshold Processing of Secondary Electron Images in Scanning Electron Microscope

Abstract: Observing the sample under a scanning electron microscope (SEM) requires adjustment of brightness and contrast to obtain a clear image. The traditional method is manually adjusted by the operator, which inevitably has errors. In this paper, an adaptive threshold processing method based on image-based normalized gray histogram is proposed. This method can acquire the threshold of the image according to the state of the currently obtained secondary electron images. When the brightness and contrast of the image change, the threshold can also be changed accordingly. It is concluded from a large number of tests that when the secondary electron image gray histogram has obvious double peaks and is located in the trough, the threshold obtained is optimal. Therefore, it is possible to better observe the pictures under the SEM.

A Review of Computer Microvision-Based Precision Motion Measurement: Principles, Characteristics, and Applications

Abstract: Microengineering/nanoengineering is an emerging field that enables engineering and scientific discoveries in the microworld. As an effective and powerful tool for automation and manipulation at small scales, precision motion measurement by computer microvision is now broadly accepted and widely used in microengineering/nanoengineering. Unlike other measurement methods, the vision-based techniques can intuitively visualize the measuring process with high interactivity, expansibility, and flexibility. This article aims to comprehensively present a survey of microvision-based motion measurement from the collective experience. Working principles of microvision systems are first introduced and described, where the hardware configuration, model calibration, and motion measurement algorithms are systematically summarized. The characteristics and performances of different microvision-based methods are then analyzed and discussed in terms of measurement resolution, range, degree of freedom, efficiency, and error sources. Recent advances of applications empowered by the developed computer microvision-based methods are also presented. The review can be helpful to researchers who engage in the development of microvision-based techniques and provide the recent state and tendency for the research community of vision-based measurement, manipulation, and automation at microscale/nanoscale.

Eosinophilia/Hypereosinophilia in the Setting of Reactive and Idiopathic Causes, Well-Defined Myeloid or Lymphoid Leukemias, or Germline Disorders.

Abstract: Objectives To report the findings of the 2019 Society for Hematopathology/European Association for Haematopathology Workshop within the categories of reactive eosinophilia, hypereosinophilic syndrome (HES), germline disorders with eosinophilia (GDE), and myeloid and lymphoid neoplasms associated with eosinophilia (excluding entities covered by other studies in this series). Methods The workshop panel reviewed 109 cases, assigned consensus diagnosis, and created diagnosis-specific sessions. Results The most frequent diagnosis was reactive eosinophilia (35), followed by acute leukemia (24). Myeloproliferative neoplasms (MPNs) received 17 submissions, including chronic eosinophilic leukemia, not otherwise specified (CEL, NOS). Myelodysplastic syndrome (MDS), MDS/MPN, and therapy-related myeloid neoplasms received 11, while GDE and HES received 12 and 11 submissions, respectively. Conclusions Hypereosinophilia and HES are defined by specific clinical and laboratory criteria. Eosinophilia is commonly reactive. An acute leukemic onset with eosinophilia may suggest core-binding factor acute myeloid leukemia, blast phase of chronic myeloid leukemia, BCR-ABL1-positive leukemia, or t(5;14) B-lymphoblastic leukemia. Eosinophilia is rare in MDS but common in MDS/MPN. CEL, NOS is a clinically aggressive MPN with eosinophilia as the dominant feature. Bone marrow morphology and cytogenetic and/or molecular clonality may distinguish CEL from HES. Molecular testing helps to better subclassify myeloid neoplasms with eosinophilia and to identify patients for targeted treatments.

A magnification-continuous calibration method for SEM-based nanorobotic manipulation systems

Abstract: Calibration for scanning electron microscope (SEM) based nanorobotic manipulation systems is important and difficult. Most current calibration methods are cumbersome because they require customized high precision calibration boards and repeated calibration procedures in different magnifications. This paper presents a convenient magnification-continuous calibration method with high precision for SEM-based nanorobotic manipulation systems. The projection matrix containing a continuous magnification factor is obtained by modifying the affine camera model. This facilitates the simplification of the parameter computing process. Movement features are used to align the moving axes of micropositioning stages and calibrate the system, which benefits for the realization of efficient automatic calibration. Three experiments are carried out, and the results demonstrate that the proposed method is effective and practical for calibrating SEM-based nanorobotic manipulation systems under a wide range of continuous magnifications. Experiments also confirm that high precision measurements can be conducted in different magnifications with only once calibration and the relative error is within 1%.

Micro-object pose estimation with sim-to-real transfer learning using small dataset

Abstract: Abstract Three-dimensional (3D) pose estimation of micro/nano-objects is essential for the implementation of automatic manipulation in micro/nano-robotic systems. However, out-of-plane pose estimation of a micro/nano-object is challenging, since the images are typically obtained in 2D using a scanning electron microscope (SEM) or an optical microscope (OM). Traditional deep learning based methods require the collection of a large amount of labeled data for model training to estimate the 3D pose of an object from a monocular image. Here we present a sim-to-real learning-to-match approach for 3D pose estimation of micro/nano-objects. Instead of collecting large training datasets, simulated data is generated to enlarge the limited experimental data obtained in practice, while the domain gap between the generated and experimental data is minimized via image translation based on a generative adversarial network (GAN) model. A learning-to-match approach is used to map the generated data and the experimental data to a low-dimensional space with the same data distribution for different pose labels, which ensures effective feature embedding. Combining the labeled data obtained from experiments and simulations, a new training dataset is constructed for robust pose estimation. The proposed method is validated with images from both SEM and OM, facilitating the development of closed-loop control of micro/nano-objects with complex shapes in micro/nano-robotic systems.