scholarly journals Detection of Fake Indian Currency Notes using Deep Learning

2021 ◽  
Vol 9 (VII) ◽  
pp. 1353-1357
Author(s):  
Soha K Deshpande
Keyword(s):  
Real Time ◽  
Identification System ◽  
Detection Accuracy ◽  
Test Results ◽  
Test Dataset ◽  
Model Based ◽  
Detection And Identification ◽  
Commercial Transaction ◽  
Normal Human ◽  
Visually Challenged

A normal human being can easily see and distinguish any banknote, however doing the same job is extremely difficult for someone who is visually challenged or blind. Because money plays such an essential part in our everyday lives and is required for any commercial transaction, real-time detection and recognition of banknotes is a must for anyone who is blind or visually impaired. The mobilenet based CNN model-based Indian currency detection and identification system is presented for this purpose, and it is quick and accurate. To make the system more resilient, pictures of various denominations and situations were collected first, and then these images were supplemented with various geometric and image modifications. These augmented pictures are then manually tagged, and training and validation image sets are created from them. Later, the trained model's performance was assessed on a real-time scenario as well as a test dataset. The suggested mobile net model-based technique exhibits detection accuracy of 91.33% according to the test results. This standalone system operates in real-time.

Robust Model Reference Fault Detection and Identification System for Fixed Wing Aircrafts

2018 ◽  
Vol 10 (5) ◽  
Author(s):  
R. Jaganraj ◽  
R. Velu
Keyword(s):  
Fault Detection ◽  
Real Time ◽  
Fault Estimation ◽  
Identification System ◽  
Model Reference ◽  
Robust Model ◽  
Detection And Identification ◽  
Fault Detection And Identification ◽  
Time Parameters ◽  
Implantable Sensor

Fault Detection and Identification system (FDI) and Fault Tolerant Flight Control (FTFC) system are used to correct the faulty operation of an aircraft. Both FDIs and FTFCs have operational disadvantages due to their inherent limitation of fault source identification. This paper presents the design and implementation of a robust model reference fault detection and identification (MRFDI) system on a fixed-wing aircraft for identifying actuator fault, instrument fault and presence of any uncertainties. The proposed MRDFI fuses the real-time parameters and actuator feedback to combine the advantages of data driven and model reference FDI that makes robust fault estimation. The MRFDI system is implemented on a typical aircraft altitude hold autopilot simulation environment with a predefined fault scenario. The fault scenario includes a faulty elevator, a faulty skin-implantable sensor and wind gust as environmental uncertainty. The MRFDI performs logical analysis to detect fault using state-dependent real-time parameters and state-independent skin implantable sensor. This two-step fault detection method makes MRFDI robust to any type of fault identification. The results show that the MRFDI detects and distinguishes faults in actuator, instrument and any of the listed uncertainties thrown by the environment accurately.

scholarly journals Raman Spectroscopic and Advanced Signal Processing analyses for Real Time Standoff Detection and Identification of Explosives

2021 ◽  
Author(s):  
Adel Abdallah ◽  
Alaaeldin Mahmoud ◽  
Mohamed Mokhtar ◽  
Aiman Mousa ◽  
Yahia Elbashar ◽  
...  
Keyword(s):  
Raman Spectroscopy ◽  
Signal Processing ◽  
Real Time ◽  
Laser Raman Spectroscopy ◽  
Identification System ◽  
Second Phase ◽  
Accurate Identification ◽  
Laser Raman ◽  
Detection And Identification ◽  
Two Phases

Abstract Laser Raman spectroscopy is a powerful instrument commonly used for detection of bulk and trace amounts of explosives. The work carried out in this paper is divided into two phases; the first phase is to propose a real time standoff explosive detection and identification system based on Raman spectroscopy that can be deployed in static checkpoints. The measurement is performed for samples placed in contact and at distances up to 1 meter in ambient light conditions. The second phase is to propose a novel sophisticated signal processing and pattern recognition techniques for accurate identification and classification of the investigated materials.

scholarly journals Using Channel and Network Layer Pruning Based on Deep Learning for Real-Time Detection of Ginger Images

Agriculture ◽  
2021 ◽  
Vol 11 (12) ◽  
pp. 1190
Author(s):  
Lifa Fang ◽  
Yanqiang Wu ◽  
Yuhua Li ◽  
Hongen Guo ◽  
Hua Zhang ◽  
...  
Keyword(s):  
Real Time ◽  
Computational Cost ◽  
Detection Accuracy ◽  
Test Results ◽  
Network Layers ◽  
Solid Foundation ◽  
Time Determination ◽  
Model Size ◽  
Detection Speed

Consistent ginger shoot orientation helps to ensure consistent ginger emergence and meet shading requirements. YOLO v3 is used to recognize ginger images in response to the current ginger seeder’s difficulty in meeting the above agronomic problems. However, it is not suitable for direct application on edge computing devices due to its high computational cost. To make the network more compact and to address the problems of low detection accuracy and long inference time, this study proposes an improved YOLO v3 model, in which some redundant channels and network layers are pruned to achieve real-time determination of ginger shoots and seeds. The test results showed that the pruned model reduced its model size by 87.2% and improved the detection speed by 85%. Meanwhile, its mean average precision (mAP) reached 98.0% for ginger shoots and seeds, only 0.1% lower than the model before pruning. Moreover, after deploying the model to the Jetson Nano, the test results showed that its mAP was 97.94%, the recognition accuracy could reach 96.7%, and detection speed could reach 20 frames·s−1. The results showed that the proposed method was feasible for real-time and accurate detection of ginger images, providing a solid foundation for automatic and accurate ginger seeding.

Improving model-based compensation method for real-time hybrid simulation considering error of identified model

2020 ◽  
pp. 107754632096162
Author(s):  
Zihao Zhou ◽  
Ning Li
Keyword(s):  
Time Delay ◽  
Transfer Function ◽  
Real Time ◽  
Hybrid Simulation ◽  
Compensation Method ◽  
Test Results ◽  
Two Stage ◽  
The Real ◽  
Model Based ◽  
Real Plant

Time delay is a critical and unavoidable problem in real-time hybrid simulation. An accurate and effective compensation method for time delay is necessary for the safety of real-time hybrid simulation and the reliability of test results. Generally, a model-based compensation method can be adopted, which is derived from the identified transfer function by assuming the latter can accurately represent the real plant. However, there must be some differences between the transfer function and the real plant. To facilitate the development of real-time hybrid simulation, we proposed a two-stage feedforward compensation method considering the error between the transfer function identified and the real plant. The compensation strategy proposed in this study was not only based on the transfer function but also introduced an error model as a second-stage compensation into a compensator to realize the synchronization of command and measurement. To verify the efficiency of the proposed method, comparisons in time domain and frequency domain with the feedforward compensator in a model-based feedforward–feedback control method were carried out. Compared with the feedforward compensator, the two-stage method achieved better tracking performance, especially in the high-frequency bandwidth. The test results verified that for a band-limited white noise of 0–30 Hz, the phase lag of the actuation system can be limited to ±5°. Finally, the two-stage method was applied to a real-time hybrid simulation of a two-story frame to illustrate its compensation effect on time delay.

COVID-19 pneumonia diagnosis using a simple 2D deep learning framework with a single chest CT image (Preprint)

2020 ◽  
Author(s):  
Jinseok Lee
Keyword(s):  
Deep Learning ◽  
Diagnostic Performance ◽  
Ct Images ◽  
Chest Ct ◽  
University Hospital ◽  
Detection Accuracy ◽  
Ct Image ◽  
Test Dataset ◽  
Learning Framework ◽  
Testing Dataset

BACKGROUND The coronavirus disease (COVID-19) has explosively spread worldwide since the beginning of 2020. According to a multinational consensus statement from the Fleischner Society, computed tomography (CT) can be used as a relevant screening tool owing to its higher sensitivity for detecting early pneumonic changes. However, physicians are extremely busy fighting COVID-19 in this era of worldwide crisis. Thus, it is crucial to accelerate the development of an artificial intelligence (AI) diagnostic tool to support physicians. OBJECTIVE We aimed to quickly develop an AI technique to diagnose COVID-19 pneumonia and differentiate it from non-COVID pneumonia and non-pneumonia diseases on CT. METHODS A simple 2D deep learning framework, named fast-track COVID-19 classification network (FCONet), was developed to diagnose COVID-19 pneumonia based on a single chest CT image. FCONet was developed by transfer learning, using one of the four state-of-art pre-trained deep learning models (VGG16, ResNet50, InceptionV3, or Xception) as a backbone. For training and testing of FCONet, we collected 3,993 chest CT images of patients with COVID-19 pneumonia, other pneumonia, and non-pneumonia diseases from Wonkwang University Hospital, Chonnam National University Hospital, and the Italian Society of Medical and Interventional Radiology public database. These CT images were split into a training and a testing set at a ratio of 8:2. For the test dataset, the diagnostic performance to diagnose COVID-19 pneumonia was compared among the four pre-trained FCONet models. In addition, we tested the FCONet models on an additional external testing dataset extracted from the embedded low-quality chest CT images of COVID-19 pneumonia in recently published papers. RESULTS Of the four pre-trained models of FCONet, the ResNet50 showed excellent diagnostic performance (sensitivity 99.58%, specificity 100%, and accuracy 99.87%) and outperformed the other three pre-trained models in testing dataset. In additional external test dataset using low-quality CT images, the detection accuracy of the ResNet50 model was the highest (96.97%), followed by Xception, InceptionV3, and VGG16 (90.71%, 89.38%, and 87.12%, respectively). CONCLUSIONS The FCONet, a simple 2D deep learning framework based on a single chest CT image, provides excellent diagnostic performance in detecting COVID-19 pneumonia. Based on our testing dataset, the ResNet50-based FCONet might be the best model, as it outperformed other FCONet models based on VGG16, Xception, and InceptionV3.

scholarly journals Power-Based Non-Intrusive Condition Monitoring for Terminal Device in Smart Grid

Sensors ◽  
2020 ◽  
Vol 20 (13) ◽  
pp. 3635 ◽  
Author(s):  
Guoming Zhang ◽  
Xiaoyu Ji ◽  
Yanjie Li ◽  
Wenyuan Xu
Keyword(s):  
Smart Grid ◽  
Power Consumption ◽  
Real Time ◽  
Condition Monitoring ◽  
Resource Constraints ◽  
Detection System ◽  
Stable Operation ◽  
Detection Accuracy ◽  
Monitoring Method ◽  
Dsp Processors

As a critical component in the smart grid, the Distribution Terminal Unit (DTU) dynamically adjusts the running status of the entire smart grid based on the collected electrical parameters to ensure the safe and stable operation of the smart grid. However, as a real-time embedded device, DTU has not only resource constraints but also specific requirements on real-time performance, thus, the traditional anomaly detection method cannot be deployed. To detect the tamper of the program running on DTU, we proposed a power-based non-intrusive condition monitoring method that collects and analyzes the power consumption of DTU using power sensors and machine learning (ML) techniques, the feasibility of this approach is that the power consumption is closely related to the executing code in CPUs, that is when the execution code is tampered with, the power consumption changes accordingly. To validate this idea, we set up a testbed based on DTU and simulated four types of imperceptible attacks that change the code running in ARM and DSP processors, respectively. We generate representative features and select lightweight ML algorithms to detect these attacks. We finally implemented the detection system on the windows and ubuntu platform and validated its effectiveness. The results show that the detection accuracy is up to 99.98% in a non-intrusive and lightweight way.

Sign in / Sign up

Export Citation Format

Share Document