With the advent of the artificial intelligence era, deep learning technology is widely used in various fields, showing a good advantage in image noise reduction. In order to let more scholars understand the progress of machine learning in the field of image noise reduction, the research on machine learning in image denoising is reviewed. This paper mainly introduces three kinds of models, such as convolutional neural network, pulse coupled neural network and wavelet neural network, which are commonly used in image denoising. The nonlocal mean noise reduction method based on machine learning is described as a concrete case. The purpose of the article is to clearly understand the latest developments in deep learning in the field of image noise reduction.

Overview of Image Denoising Based on Deep Learning

The power window control is the mechanism to automatically control the upward and downward movement of power window in vehicles by application of position, current, flexi force and temperature sensor as replacement of conventional or mechanical controlled hand crank based system. The paper focuses on model based development and testing of miniaturized scale controller for automotive smart power window, hardware-in-the-loop (HIL) testing using dSPACE simulator. The experimental work is carried out for the same to provide the desired precision due to the testing limitation in real time vehicle. The HIL set up is utilized to analyze the different manual and sensory inputs functionalities in Indian driving environment. The paper presented the DC motor interfacing to the controller as the main building block of the power window system. The motor takes the decision based on the decision tree classifier (DTC) machine learning algorithm and control movements in desired direction by condition inputs from decision tree. The system behavior is completely based on the algorithm, dSPACE testing environment supports the window functionality and its movement control in upward, downward direction and the results showed that the developed system is effective in identifying obstacle. The decision tree test set and training set data depicts that 93%-sampled results are valid outcome and 7% are invalid under 12 different test observations.

Automotive Power Window Communication with DTC Algorithm and Hardware-in-the Loop Testing

The research paper emphasizes on the design and real time implementation of microcontroller based power window control used as a control system for moving a power window panel. The purpose of power window control system is to raise and lower door glass with the help of a switch and its operation is controlled by the use of following sensors LPG sensor, LDR sensor and position sensor to replace the use of hand-turned crank handle. Power window system consists of driver motor, power electronic and control system. The control unit senses both hard and soft obstructions and deactivates a motor that moves the glass when any obstruction is detected. The controller also senses obstructions on start-up of the motor and regulates the speed of the window panel by pulse width modulating motor signals. In other way the controller senses obstructions by maintaining data which is related to motor operation in three different memory buffers that are regularly updated upon receipt of pulses that are related to motor speed. The whole system can be designed as an intelligent control system by applying number of conditions which results to the movement of the power window. It increases the safety of the automobile as well as increases human comfort inside the vehicle. Such applications are fire safety in the vehicle, over temperature detection, sensing of LPG leakage. Hardware implementation includes some basic input switches, sensors connected to the microcontroller and the LEDs and LCD are used to show the status of output. Input sensors used for this research are LPG sensor, LDR sensor and position sensor, output is the functioning of the motor. The implementation work is carried out using AVR ATMEGA 16 microcontroller.

Design and hardware development of power window control mechanism using microcontroller

A close relation between gravitational waveforms and the types of trajectories in a superposed field between a pseudo-Newtonian Kerr black hole and quadrupolar halos is shown in detail. The gravitational waveforms emitted from circular, KAM tori and chaotic orbits must be periodic, quasiperiodic and stochastic, respectively. The chaotic motion can maximally enhance both the amplitudes and the energy emission rates of the waves.

http://iopscience.iop.org/article/10.1088/0253-6102/56/6/13/pdf

Gravitational Waves from a Pseudo-Newtonian Kerr Field with Halos

Non-Local Means (NLM) denoising algorithm uses Gaussian weighted Euclidean distance to measure the similarity between neighborhoods, but when the image noise level is high, the distance measurement method is not well filtered noise. In this paper, an improved nonlocal mean filter image denoising algorithm is designed by analyzing the shortcomings of Gaussian weighted Euclidean distance in measuring neighborhood similarity. The gradient information of a filtered image is used and applied to the Gaussian weighting coefficient to achieve the purpose of adaptive filtering. The experimental results show that the algorithm has better denoising effect than traditional algorithm, especially when the noise level is high, it has high peak signal to noise ratio.

An improved non-local means algorithm for image denoising

Aiming at the problem of the over-smoothness and blurs the details,which are caused by exponential kernel function used in original non-local means algorithm,this paper proposes a cosine Gaussian kernel function based on exponential kernel function and combined with a cosine coefficient and Gaussian kernelIt is used in the weight-computing of the improved algorithmExperimental results show the algorithm has a superior denoising performance than the original one,especially with detail information in the image,and PSNR can be improved by 16 dB at most

Improved Non-local Means Algorithm for Image Denoising

A design of car window antitrapping without sensor is proposed,which is insalled in the window frame.The DC motor is drived by the relays,the current of the motor is processed by a PIC MCU through an A/D converter to judge whether the car window has bumped a barrier.In this way the function of antitrapping has been implemented.

Design of Car Window Antitrapping Without Sensor

This paper focuses on the relic gravity waves produced during the transition from a radiation-dominated inflationary phase to a dust-dominated Friedman–Robertson–Walker-type expansion. We discuss how to investigate the spectral energy density by the latest space-based GWs detectors at f = 0.1 Hz (i.e. DECIGO). In the case of power-law and exponential inflation, we apply the cross-correlation method to the latest detector and get the time dependence of the very early Hubble parameter.

Relic Gravity Waves Investigation by Advanced Space-Based Gravitational Waves Detector

This paper discusses the gravitation waveforms of binaries coalescence into intermediate mass black holes (about 30 times of the solar mass). We focus on the non-spinning intermediate mass black hole located less than 100 Mpc from earth. By comparing two simulation waveforms (effective one body numerical relativity waveform (EOBNR), phenomenological waveform), we discuss the relationship between the effective distance and frequency; and through analyzing large amounts of data in event, we find that the phenomenological waveform is much smoother than EOBNR waveform, and has higher accuracy at the same effective distance.

http://iopscience.iop.org/article/10.1088/0253-6102/57/1/09/pdf

Zhong Yuan-hong

Papers

Improved Non-local Means Algorithm for Image Denoising

Design of Car Window Antitrapping Without Sensor

Relic Gravity Waves Investigation by Advanced Space-Based Gravitational Waves Detector

Gravitational Radiation of Binaries Coalescence into Intermediate Mass Black Holes