Although many autofocus (AF) algorithms have been proposed and compared in the literature [1], [2], none of the existing algorithms can work perfectly for images of a scanning electron microscope (SEM) in practice. This is because a simple mathematical scalar metric cannot perfectly capture the quality of images, especially for a variety of SEM samples, hardware specifications, measurement environment, and etc. In addition, a simple scalar AF metric cannot govern a variety of control parameters such as sharpness, contrast, and brightness. At the era of the 4th industrial revolution, the ultimate goal would be a fully autonomous machine controlling the SEM parameters to get the best image just as a SEM specialist does. To take the first step to the automatic machine-controlled SEM development, we propose a supervised learning framework that automatically assesses the quality of sample images as if a SEM specialist does. Specifically, we develop a deep learning computer software that uses an input of a sample image and current control parameters such as brightness, contrast, and focus to automatically score the quality of the sample image. To evaluate how accurately the proposed deep learning software can score, we define a mean squared error loss (MSE) L as 1 2 1 , where is the score that a SEM specialist gives and is the score predicted by the proposed software. The following neural network architectures are used for deep learning: i) parallelized convolutional neural network (CNN) and fully-connected neural network (FCNN) (PACF) as in Fig. 1, ii) VGG [3], and iii) ResNet [4]. Gray-scale images with a sample type of a grid or tin ball are used. The original image resolution is 640x480-pixel; however, we augment the amount of data by cropping the images to 224x224 pixels, flipping them vertically, and rotating them randomly. As a result, the total number of images is 2134. For supervised learning, the deep learning network should be trained with known inputs and outputs. Thus, 1493 images out of the 2134 images are used for training, and the remaining 641 images are used for testing. Control parameters such as brightness, contrast, and focus of each image are set differently. In addition, each image is scored by a SEM specialist as an overall score on the quality of the image. Experiments are implemented under the following environment: GeForce GTX 1080M, CUDA 10.0, Python 3.6.7, Pytorch 1.0.0, and OpenCV 4.0.1. Table 1 shows the performance comparison for the average test runtime of each image and the Root of MSE (RMSE) between the existing AF algorithms and deep-learning based score prediction networks for 100 epoch and 0.00015 learning rate. For comparison, the values of the AF functions are normalized to scores by . The proposed networks outperform the existing AF functions with respect to both of the test runtime and RMSE. Fig. 2 shows the examples of the scores from the specialist, one of the existing AF algorithms based on absolute variance, and one of the proposed networks, PACF, respectively. Even in the ...

No abstract

This paper presents a wearable wireless surface electromyogram (sEMG) integrated interface that utilizes a proposed analog pseudo-wavelet preprocessor (APWP) for signal acquisition and pattern recognition. The APWP is integrated into a readout integrated circuit (ROIC), which is fabricated in a 0.18-µm complementary metal-oxide-semiconductor (CMOS) process. Based on this ROIC, a wearable device module and its wireless system prototype are implemented to recognize five kinds of real-time handgesture motions, where the power consumption is further reduced by adopting low-power components. Real-time measurements of sEMG signals and APWP data through this wearable interface are wirelessly transferred to a laptop or a sensor hub, and then they are further processed to implement the pseudo-wavelet transform under the MATLAB environment. The resulting APWP-augmented pattern-recognition algorithm was experimentally verified to improve the accuracy by 7 % with a real-time frequency analysis.

A joint resource allocation (RA), user association (UA), and power control (PC) problem is addressed for proportional fairness maximization in a cooperative multiuser downlink small cell network with limited backhaul capacity, based on orthogonal frequency division multiplexing. Previous studies have relaxed the per-resourceblock (RB) RA and UA problem to a continuous optimisation problem based on long-term signal-to-noise-ratio, because the original problem is known as a combinatorial NP-hard problem. We tackle the original per-RB RA and UA problem to obtain a near-optimal solution with feasible complexity. We show that the conventional dual problem approach for RA cannot find the solution satisfying the conventional KKT conditions. Inspired by the dual problem approach, however, we derive the first order optimality conditions for the considered RA, UA, and PC problem, and propose a sequential optimization method for finding the solution. The overall proposed scheme can be implemented with feasible complexity even with a large number of system parameters. Numerical results show that the proposed scheme achieves the proportional fairness close to its outer bound with unlimited backhaul capacity in the low backhaul capacity regime and to that of a carefully-designed genetic algorithm with excessive generations but without backhaul constraint in the high backhaul capacity regime. Index TermsOrthogonal frequency division multiplexing (OFDM) downlink, proportional fairness maximization, user association, resource allocation, power control, cooperative small cells network, limited backhaul capacity This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible. ). I. INTRODUCTIONIt has long been a challenge to suppress intercell interference in cellular mobile networks, and thereby improving the system throughput and spectral efficiency. Indeed, it is known that the cell densification may degrade the sum-rate unless the level of intercell interference is kept low enough compared to the desired signal [1]. The concept of "small cells" is one of the enablers of the next generation mobile network requiring extremely high data rate connections, where multiple small cell base stations (SBSs) in proximity are clustered to make a hotspot area providing high data rate connectivity. As the SBS cell size becomes smaller to further increase the sum-rate, user association (UA), resource allocation (RA), and power control (PC) should be carefully designed to mitigate intercell interference, particularly for cell-edge users. The optimisation for UA, RA, and PC can be considered to maximize the fairness of the users [2] or the sum-capacity of the total system [3]- [6]. The focus of this paper is on the proportional fairness maximization, where the aim is to maximize the geometric mean of users' rates, compromising between the fairness and sum-capacity maximization.A variety of literature have been around to tackle the joint optim...

This is an Accepted Manuscript for the Microscopy and Microanalysis 2020 Proceedings. This version may be subject to change during the production process.