Diabetes leads to health problems for hundreds of millions of people globally every year. Available medical records of patients quantify symptoms, body features, and clinical laboratory test values, which can be used to perform biostatistics analysis aimed at finding patterns or features undetectable by current practice. In this work, we proposed a machine learning model to predict the early onset of diabetes patients. It is a novel wrapper-based feature selection utilizing Grey Wolf Optimization (GWO) and an Adaptive Particle Swam Optimization (APSO) to optimize the Multilayer Perceptron (MLP) to reduce the number of required input attributes. Moreover, we also compared the results achieved using this method and several conventional machine learning algorithms approaches such as Support Vector Machine (SVM), Decision Tree (DT), K-Nearest Neighbor (KNN), Naïve Bayesian Classifier (NBC), Random Forest Classifier (RFC), Logistic Regression (LR). Computational results of our proposed method show not only that much fewer features are needed, but also higher prediction accuracy can be achieved (96% for GWO -MLP and 97% for APGWO -MLP). This work has the potential to be applicable to clinical practice and become a supporting tool for doctors/physicians.

Plant disease, especially crop plants, is a major threat to global food security since many diseases directly affect the quality of the fruits, grains, and so on, leading to a decrease in agricultural productivity. Farmers have to observe and determine whether a leaf was infected by naked eyes. This process is unreliable, inconsistent, and error prone. Several works on deep learning techniques for detecting leaf diseases had been proposed. Most of them built their models based on limited resolution images using convolutional neural networks (CNNs). In this research, we aim at detecting early disease on plant leaves with small disease blobs, which can only be detected with higher resolution images, by an artificial neural network (ANN) approach. After a pre-processing step using a contrast enhancement method, all the infested blobs are segmented for the whole dataset. A list of several measurement-based features that represents the blobs are chosen and then selected based on their influences on the model's performance using a wrapperbased feature selection algorithm, which is built based on a hybrid metaheuristic. The chosen features are used as inputs for an ANN. We compare the results obtained using our methods with another approach using popular CNN models (AlexNet, VGG16, ResNet-50) enhanced with transfer learning. The ANN's results are better than those of CNNs using a simpler network structure (89.41% vs 78.64%, 79.92%, and 84.88%, respectively). This shows that our approach can be implemented on low-end devices such as smartphones, which will be of great assistance to farmers on the field. INDEX TERMS Neural network, image classification, plant disease, feature selection, precision agriculture.

In 2002, we reported a CCD image sensor with 260×312 pixels capable of capturing 103 consecutive images at 1,000,000 frames per second (1Mfps) [1]. We named the sensor "ISIS-V2", for In-situ Storage Image Sensor Version 2. 103 memory elements are attached to every pixel; generated image signals were instantly and continuously stored in the in-situ storage without being read out of the sensor. The ultimate high-speed recording was enabled by this parallel recording at all pixels. In 2006, the color version, ISIS-V4, was reported [2]. In 2009, we developed ISIS-V12, a backside-illuminated image sensor mounting the ISIS structure and the CCM, charge-carrier multiplication, on the front side [3]. The CCM is a CCD-specific efficient signal-amplification device. CCM, combined with the BSI structure and cooling, achieved very high sensitivity. The ISIS-V12 was a test sensor intended to prove the technical feasibility of the structure. The maximum frame rate was 250kfps for a charge-handling capacity of Q max =10,000e -and 1Mfps for a reduced Q max . The pixel count was 489×400 pixels. For backside-illuminated (BSI) image sensors, metal wires can be placed on the front surface to increase the frame rate without reducing fill factor or violating uniformity of the pixel configuration. It has been proved by simulations that 100Mfps is achievable by introducing innovative technologies including a special wiring method [4]. We now report on ISIS-V16, developed by incorporating technologies to increase the frame rate with those to achieve very high sensitivity, which was confirmed by evaluation of ISIS-V12. The performance specification of ISIS-V16 is summarized in Fig. 23.4.1. Figure 23.4.2 shows the global planar structure of ISIS-V16. The imaging area is divided into 4 rectangular subareas. A set of driving voltages used in the image-capturing operation, which requires very high frequency, is transferred from the left and right, toward the vertical center-line through metal inner bus lines. The very wide inner bus lines, significantly reduce the resistance. The inner bus lines are connected to the outer metal bus lines with a special shape, named "Thunderbolt bus lines," which also serve to reduce the resistance in transferring the driving voltages.Figs. 23.4.3 and 23.4.4 depict the plane structure, installed on the front side, and a cross-section taken along the A-A' line in Fig. 23.4.3. In Fig. 23.4.4, incident photons generate electron-hole pairs in the thick p -generation layer. The generated electrons travel to the collection gate on the front side to form a signal charge packet. The charge packet is then transferred along an n + CCD channel, which is a memory device extending linearly in a slightly slanted direction to the orthogonal direction to the sheet, as shown in Fig. 23.4.3.In Fig. 23.4.3, a signal charge packet is transferred from the collection gate to the memory CCD channel, carried downward and drained from the drain at the end of the CCD channel. Therefore, a sequence of the latest image signals is always ...

Among the physical attributes of agricultural materials, mass, volume, and sizes have always been important quality parameters. Previous research focused mostly on volume estimation using stereo-based approaches, which rely on manual intervention or require a multiple-cameras set up or multiple-frames captures from different viewing angles to reconstruct the three-dimensional point-cloud information. These approaches are tedious and not suitable for practical machine vision systems. In this work, we only use a single camera mounted on the ceiling of the imaging chamber, which is directly above the fruit/vegetable to capture its top-view, two-dimensional image. We developed a method to estimate the mass/volume of agricultural products with axi-symmetrical shapes such as a carrot or a cucumber. The mass/volume is estimated as the sum of smaller standard blocks, such as chopped pyramids, an elliptical cone, or a conical cone. The computed mass/volume showed good agreement with analytical and experimental results. The weight estimation error is 95% for the case of the carrot and 96.7% for the cucumber. The method proved to be sufficiently accurate, easy to use, and rotationally invariant.

Parkinson’s Disease (PD) is a brain disorder that causes uncontrollable movements. According to estimation, roughly ten million individuals worldwide have had or are developing PD. This disorder can have severe consequences that affect the patient’s daily life. Therefore, several previous works have worked on PD detection. Automatic Parkinson’s Disease detection in voice recordings can be an innovation compared to other costly methods of ruling out examinations since the nature of this disease is unpredictable and non-curable. Analyzing the collected vocal records will detect essential patterns, and timely recommendations on appropriate treatments will be extremely helpful. This research proposed a machine learning-based approach for classifying healthy people from people with the disease utilizing Grey Wolf Optimization (GWO) for feature selection, along with Light Gradient Boosted Machine (LGBM) to optimize the model performance. The proposed method shows highly competitive results and has the ability to be developed further and implemented in a real-world setting.