Transfer learning based handwritten character recognition of tamil script using inception-V3 Model

The paper describes the excellent method to get first-rate accuracy and performance in the discipline of Tamil character recognition in a handwritten mode. However, the subject is still at a nascent stage and grossly lacks adequate accuracy in the Tamil language, even though several studies have been conducted within the discipline of handwritten character recognition. This paper draws the attention to the offline handwritten recognition for the Tamil language using the Inception-v3 based transfer learning method. The proposed work is conducted on the readily available HP Tamil handwritten character offline dataset (Hewlett-Packard Lab: hpl-tamil-iso-char-offline-1.0.). It reveals that with the suitable arrangement of transfer learning approach with Inception-v3, the pre-trained model can achieve the recognition accuracy of 93.1%, overtaking the former deep learning designs. The achieved accuracy is due to the use of a pre-trained version with transfer learning that regularly hastens the method of the training process on a new task. Overall, this results in higher accuracy and a more capable version.

Mallard Detection Using Microphone Arrays Combined with Delay-and-Sum Beamforming for Smart and Remote Rice–Duck Farming

This paper presents an estimation method for a sound source of pre-recorded mallard calls from acoustic information using two microphone arrays combined with delay-and-sum beamforming. Rice farming using mallards saves labor because mallards work instead of farmers. Nevertheless, the number of mallards declines when they are preyed upon by natural enemies such as crows, kites, and weasels. We consider that efficient management can be achieved by locating and identifying the locations of mallards and their natural enemies using acoustic information that can be widely sensed in a paddy field. For this study, we developed a prototype system that comprises two sets of microphone arrays. We used 64 microphones in all installed on our originally designed and assembled sensor mounts. We obtained three acoustic datasets in an outdoor environment for our benchmark evaluation. The experimentally obtained results demonstrated that the proposed system provides adequate accuracy for application to rice–duck farming.

Smart Pothole Detection Using Deep Learning Based on Dilated Convolution

Roads make a huge contribution to the economy and act as a platform for transportation. Potholes in roads are one of the major concerns in transportation infrastructure. A lot of research has proposed using computer vision techniques to automate pothole detection that include a wide range of image processing and object detection algorithms. There is a need to automate the pothole detection process with adequate accuracy and speed and implement the process easily and with low setup cost. In this paper, we have developed efficient deep learning convolution neural networks (CNNs) to detect potholes in real-time with adequate accuracy. To reduce the computational cost and improve the training results, this paper proposes a modified VGG16 (MVGG16) network by removing some convolution layers and using different dilation rates. Moreover, this paper uses the MVGG16 as a backbone network for the Faster R-CNN. In addition, this work compares the performance of YOLOv5 (Large (Yl), Medium (Ym), and Small (Ys)) models with ResNet101 backbone and Faster R-CNN with ResNet50(FPN), VGG16, MobileNetV2, InceptionV3, and MVGG16 backbones. The experimental results show that the Ys model is more applicable for real-time pothole detection because of its speed. In addition, using the MVGG16 network as the backbone of the Faster R-CNN provides better mean precision and shorter inference time than using VGG16, InceptionV3, or MobilNetV2 backbones. The proposed MVGG16 succeeds in balancing the pothole detection accuracy and speed.

A novel method for in-cylinder pressure prediction using the engine instantaneous crankshaft torque

Current practices of condition assessment in large marine engines are largely based on the measurement of cylinder pressure using external kits, which poses challenges due to sensors synchronisation and durability issues, as well as the inability to perform continuous monitoring. For addressing these challenges, this study aims at developing a novel method to solve the inverse problem of predicting the pressure variations in all engine cylinders, by using the Instantaneous Crankshaft Torque (ICT) measurement for large internal combustion engines. This method is developed by considering the Initial Value Problem (IVP) technique along with the integration of a direct crankshaft dynamics model incorporating the sensitivity parameters and stability criteria calculation based on the Lyapunov Exponent (LE) as well as a state-of-the-art Nonmonotone Self-Adaptive Levenberg-Marquardt (NSALMN) optimisation algorithm. The method is tested for a number of case studies using different combustion models based on the Weibe and sigmoid functions, as well as for healthy, degraded and faulty engine conditions. The derived results demonstrate adequate accuracy exhibiting a maximum error of 0.3% in the prediction of the mean peak in-cylinder pressure. The analysis of the calculated sensitivity parameters resulted in the identification of the parameters that significantly impact the solution, thus providing improved insights for selecting the developed method settings. The developed method renders the continuous and non-intrusive in-cylinder pressures monitoring feasible, by using a permanently installed shaft power metre sensor with higher sample rates.

DistoX calibration tools and the need for calibration checking

For proper cave surveying using DistoX, the device needs to be calibrated with adequate accuracy. Calibrating does not require any tools; but, tools to make calibration easier have been developed. Theoretical consideration shows that the use of certain tools enables one to introduce a type of calibration error that goes undetected by the calibration software. In this study, the existence of such errors is experimentally confirmed and their magnitude is estimated. It is demonstrated to be crucial that the DistoX is calibrated and that the calibration is valid, that is, that the device has not changed since it was last calibrated. No part of the DistoX must have moved or changed its magnetization since calibration, not even the battery. The calibration method used and the quality of the resulting calibration are important too. It is highly recommended that the DistoX be checked immediately before surveying a cave and thus avoid the possibility of using an uncalibrated, not validly calibrated, or poorly calibrated device. To complete the check, a few survey shots are measured multiple times with the device at different roll angles, and the back shot of one of the shots is measured. If the device is properly calibrated, the measurements will agree with each other within the acceptable measurement error. This is not the case for a device that is not properly calibrated.

Load Forecasting Based on LSTM Neural Network and Applicable to Loads of “Replacement of Coal with Electricity”

With the complete implementation of the “Replacement of Coal with Electricity” policy, electric loads borne by urban power systems have achieved explosive growth. The traditional load forecasting method based on “similar days” only applies to the power systems with stable load levels and fails to show adequate accuracy. Therefore, a novel load forecasting approach based on long short-term memory (LSTM) was proposed in this paper. The structure of LSTM and the procedure are introduced firstly. The following factors have been fully considered in this model: time-series characteristics of electric loads; weather, temperature, and wind force. In addition, an experimental verification was performed for “Replacement of Coal with Electricity” data. The accuracy of load forecasting was elevated from 83.2 to 95%. The results indicate that the model promptly and accurately reveals the load capacity of grid power systems in the real application, which has proved instrumental to early warning and emergency management of power system faults.

Catalytic Conversion of Xylose to Furfural by p-Toluenesulfonic Acid (pTSA) and Chlorides: Process Optimization and Kinetic Modeling

Furfural is one of the most promising precursor chemicals with an extended range of downstream derivatives. In this work, conversion of xylose to produce furfural was performed by employing p-toluenesulfonic acid (pTSA) as a catalyst in DMSO medium at moderate temperature and atmospheric pressure. The production process was optimized based on kinetic modeling of xylose conversion to furfural alongwith simultaneous formation of humin from xylose and furfural. The synergetic effects of organic acids and Lewis acids were investigated. Results showed that the catalyst pTSA-CrCl3·6H2O was a promising combined catalyst due to the high furfural yield (53.10%) at a moderate temperature of 120 °C. Observed kinetic modeling illustrated that the condensation of furfural in the DMSO solvent medium actually could be neglected. The established model was found to be satisfactory and could be well applied for process simulation and optimization with adequate accuracy. The estimated values of activation energies for xylose dehydration, condensation of xylose, and furfural to humin were 81.80, 66.50, and 93.02 kJ/mol, respectively.

Plant Leaf Disease Recognition Using Depth-Wise Separable Convolution-Based Models

Proper plant leaf disease (PLD) detection is challenging in complex backgrounds and under different capture conditions. For this reason, initially, modified adaptive centroid-based segmentation (ACS) is used to trace the proper region of interest (ROI). Automatic initialization of the number of clusters (K) using modified ACS before recognition increases tracing ROI’s scalability even for symmetrical features in various plants. Besides, convolutional neural network (CNN)-based PLD recognition models achieve adequate accuracy to some extent. However, memory requirements (large-scaled parameters) and the high computational cost of CNN-based PLD models are burning issues for the memory restricted mobile and IoT-based devices. Therefore, after tracing ROIs, three proposed depth-wise separable convolutional PLD (DSCPLD) models, such as segmented modified DSCPLD (S-modified MobileNet), segmented reduced DSCPLD (S-reduced MobileNet), and segmented extended DSCPLD (S-extended MobileNet), are utilized to represent the constructive trade-off among accuracy, model size, and computational latency. Moreover, we have compared our proposed DSCPLD recognition models with state-of-the-art models, such as MobileNet, VGG16, VGG19, and AlexNet. Among segmented-based DSCPLD models, S-modified MobileNet achieves the best accuracy of 99.55% and F1-sore of 97.07%. Besides, we have simulated our DSCPLD models using both full plant leaf images and segmented plant leaf images and conclude that, after using modified ACS, all models increase their accuracy and F1-score. Furthermore, a new plant leaf dataset containing 6580 images of eight plants was used to experiment with several depth-wise separable convolution models.

Automated and Robust Organ Segmentation for 3D-based Internal dose Calculation

Purpose: In this work we address image segmentation within dosimetry using deep learning and make three main contributions: a) to extend and op- timize the architecture of an existing Convolutional Neural Network (CNN) in order to obtain a fast, robust and accurate Computed Tomography (CT) based organ segmentation method for kidneys and livers; b) to train the CNN with an inhomogeneous set of CT scans and validate the CNN for daily dosimetry; c) to evaluate dosimetry results obtained using automated organ segmentation in comparison to manual segmentation done by two independent experts. Methods: We adapted a performant deep learning approach using CT-images to calculate organ boundaries with sufficiently high and adequate accuracy and processing time. The segmented organs were consequently used as binary masks for further convolution with a point spread function to retrieve the ac- tivity values from quantitatively reconstructed SPECT images for ”volumet- ric”/3D dosimetry. The retrieved activities were used to perform dosimetry calculations considering the kidneys as source organ. Results: The computational expenses of the algorithm was adequate enough to be used in clinical daily routine, required minimum pre-processing and per- formed within an acceptable accuracy of 93 . 4% for liver segmentation and of 94 . 1% for kidney segmentation. Additionally, kidney self-absorbed doses calcu- lated using automated segmentation differed 6 . 3% from dosimetries performed by two medical physicists in 8 patients. Conclusion: The proposed approach may accelerate volumetric dosimetry of kidneys in molecular radiotherapy with 177Lu-labelled radio-pharmaceuticals such as 177Lu-DOTATOC. However, even though a fully automated segmen- tation methodology based on CT images accelerates the organ segmentation and performs with high accuracy, it does not remove the need for supervision and corrections by experts, mostly due to misalignments in the co-registration between SPECT and CT images.Trial registration: EudraCT, 2016-001897-13. Registered 26.04.2016, www.clinicaltrialsregister.eu/ctr-search/search?query=2016-001897-13

How to Estimate the Binder: Aggregate Ratio From Lime-Based Historic Mortars for Restoration?

One of the most important studies from historic mortars is the binder:aggregate ratio, which is usually obtained through wet chemical analysis. Instrumental techniques and benchtop equipment have become increasingly important tools in the characterization of historic materials. The analysis of such materials has become more practical, faster and more accurate, and the sample preparation methods require less and less material. Thus, this article aims to investigate the validity of the results obtained by some of the methods and techniques used in historic materials analysis and determine the possibility of estimating the binder:aggregate ratio with adequate accuracy and precision. For this purpose, historic mortars from Belém do Pará, in northern Brazil, were selected, and the following quantification techniques were employed: wet chemical analysis, XRD, DSC and XRF. The results showed that the amounts of the components in the mortars could be quantified with the use of approximately 3 g of sample, thus providing one of the main pieces of information needed for the production of a restoration mortar: the binder:aggregate ratio.