@cmrit.ac.in
Associate Professor
CMR Institute of Technology
VLSI, Image processing
Scopus Publications
Scholar Citations
Scholar h-index
R. Bindu, M. Preethi Sejal, and H. Chetan
Springer Nature Singapore
T Rashmi, Anushka Jemima, C Raghavendran, and H Chetan
IEEE
Speech is an easy-to-understand communication medium that also acts as an interface for processes in artificial intelligence. In this paper, we provide a brief introduction to the humanoid that we have developed. Our humanoid is based on voice recognition principles. Voice commands are given to the humanoid as a set of instructions. The Raspberry Pi 4 microcontroller plays a significant role in collecting samples as Pulse width modulation (PWM). The signals are then transferred to the raspberry pi via a USB module. Raspberry pi-4 functions as the brain of the robot and converts the signals into a set of commands that the robot will then execute. A Servo Motor (MG90) is used to control the humanoid’s movement. We measure the success of the robot based on the accuracy of speech recognition using Google speech API and have provided the results through this paper.
H. Chetan, S. Praveen, S. Shreyas, Samridhi Singh, and R. Urvi
Springer Singapore
Meghana Nagesha, H. Chetan, and Prasad Borannavar
Springer Singapore
Kavitha. V and Chetan H
This paper proposes facial expression recognition system developed through machine learning which is capable of recognizing the emotions in real time. The work is divided mainly into two parts. In the first part the face in the image is recognized using Haar cascade classifier and AdaBoost algorithm, the idea introduced by Viola and Jones. This method can be adapted to identify the facial features rapidly with high accuracy rate. In real time application the model is capable of computing 15 frames per second. The second part deals with recognizing the emotions based on facial features identified in the first part. The system classifies number of faces simultaneously in real time. This model is implemented with deep learning algorithms using Convolutional Neural Network. Neural network is trained using the dataset FER2013 to identify different emotions. The task is to attribute emotional states such as happy, sad etc. to subjects based on their facial characteristics. Extracting the emotions from one’s profile facilitate the machines to interact with the public in various ways aiding in the development of new innovative applications. Machine learning and deep neural network acquire the information from complex appearance and categorize the patterns.
Jyoti M Roogi, Chetan H, Vinayaka R Karanji, Roshan C Thomas, Sagar T H, and Praveen Yadav
IEEE
This research work propsoes an architecture to generate Gold codes for ultrahigh speed applications using LFSRs. To generate pseudo random sequences, XOR of these sequences is obtained by using CMOS differential switch to produce gold codes. Differential switch is also used at LFSR level to XOR the tapped bits. This allows XOR feedback circuits like Gold Code Generator to operate at very high frequencies in the range of gigabits. This work deals with XOR gate implementation by using a CMOS differential switch. The results of four-input XOR gate shows that the delay is much improved.
Mahesh S Gour, Druva Kumar S, Pradeep Kumara, Manjunatha S, Sunil Kumar K, and Chetan H
IEEE
Road accidents are becoming very common in the country. The impact of road accidents can lead to the loss of many lives and can also damage many body parts. This situation becomes more serious if the riders won't wear the helmet which can be prevented by wearing the helmet and can reduce these impacts. While riding the bike, the government made it a mandatory rule to wear the helmet. Using this rule as a base, a smart helmet system is proposed which helps in providing safety to the riders and prevents accidents. The system mainly consists of Arduino Uno as a processor for processing the data, GSM & GPS modules for tracking the location and sending a message to authorized numbers, a wiper for wiping the raindrops on the helmet screen, a vibration sensor for alerting, in case the rider meets an accident and alcohol sensors as breath analyzer for the rider. The system will ensure a safe journey for riders and gives a helping hand in case of emergency. The cost of installing the whole system onto the helmet is affordable.
Rashmi T, Geethanjali S, and Chetan H
IEEE
Lung cancer is a condition which causes uncontrollable division of cells in the lungs. This causes tumor growth, which decreases a person's breathing ability. There are two types of Lung cancer - small and non-small. Both these forms of lung cancer behave differently, and have different treatments. Nevertheless, it can be difficult to identify lung cancer in its earliest stages, because the symptoms may be close to those of a respiratory infection or there may be no symptoms at all. The main aim of this paper is to find different ways to identify lung tumors using methods of photonics. We here discuss a few methods with its pros and cons. Among these we find few methods which are effective when compared to the others on basis of less time consumption, specificity, sensitivity and precision. These are discussed in detail in our discussion section. We can thereby infer that the future study of these approaches would help to diagnose and cure lung cancer in earlier stages.
Sunil Kumar K H, Chetan H, and Indumathi G
IEEE
Limitations in storage capacity and communication bandwidth made researchers to work more on image compression. Many algorithms have been implemented to compress an image for better compression ratio. Recently the application of neural networks to compress image increasing gradually and achieved great success. The proposed work uses Recurrent Neural Networks (RNN) based encoding scheme in wavelet domain transform. Dual-Tree Complex Wavelet Transform (DTCWT) uses three level tree structured Discrete Wavelet Transform (DWT) used to get eight frequency subbands, the lower sub-band output is given to RNN for final encoding process. The experimental results show that the proposed work achieves better compression ratio for higher pixel values as compared to state-of-the algorithms. The Compression Ratio (CR) can be increased for lower pixel values by using more number of iterations for the neural networks and also by modifying the wavelet components.
Anushka Jemima, C Raghavendran, and Chetan H
IEEE
Robots are characterized as soft or hard based on the material used. While hard robots are made up of materials like steel hence have high stiffness to prevent deformation and have actuators placed at the joint's soft robots, on the other hand, soft robots have soft deformable bodies made up of unconventional materials and provide a large degree of freedom. In this paper we focus on the different methods used to implement a soft robot. The main focus is laid on the fact that soft robots draw inspiration from bio-systems the classic example of an octopus arm is explained in brief. The flexibility and varying degrees of freedom as seen in animals and plants (biological systems) are brought about by the use of smart materials, soft actuators and sensors. This paper provides a brief overview of the various actuators and materials that are available, the methods of implementing them and their advantages and disadvantages. In conclusion we provide methods that are best suited for the implementation of a soft robot based on available reports and studies conducted on them. The main focus of a soft robot is to provide good contact between human and machine, wearable technology adaptability, easy handling and so on this can be achieved by the proper integration of the right materials and actuators.
N Chandra Shekar, B S Rakshitha, N Mamatha, H Chetan, M R Jyoti, and S Suganya
IOP Publishing
N Meghana and H Chetan
IEEE
The intrusion of networks has increased enormously and become major threat to both the public platform and for the classified platforms used by private/government bodies. In order to stay out from the sight of intruder a new mosaic image technique is introduced in this research work, which aims at bringing a high amount of security for the data that is being transmitted over the internet. A method is suggested in this paper where two images are merged to form a single image. The resulting image which needs to be transmitted is called as the secret image. This is then converted into tile format and is made to hide behind the image blocks. A target image is chosen which is similar to the image that is to be transmitted. While splitting the two images the different attributes like Colour conversion, generation of keys, obtaining similar blocks etc. needs to be considered carefully. For the process of encryption and decryption, a secret ECC key is used. The proposed results, is then compared with the other algorithms like RSA, AES etc. The solution obtained using ECC key method is encouraging. The use of ECC key increases system performance because of the reduced key size and thus, helps to encrypt and decrypt an image in an easier and quicker manner.
The paper describes the analysis of FFT, DWT and DTCDWT for effective transmission of acquired underwater sensor data in AWGN and Acoustic channel. The underwater temperature is acquired using a temperature sensor, pH value is acquired using pH sensor, and Depth of water is acquired using ultrasonic sensor. The sensor data is acquired using ARM Cortex M4 Microcontroller and the same data is processed using FFT, DWT and DTCDWT in AWGN channel and Acoustic channel for comparative analysis. Number of Samples is limited to 64 samples and SNR values ranging from 0 to 50 are considered for analysis. Graphical user interface is designed using MATLAB and the simulation results shows that number of errors after transmission in FFT is more compared to DWT and also number of errors is decreased in the case of DTCDWT.
The paper describes the design of underwater acoustic channel model for OFDM communication system. The underwater acoustic channel considered includes multipath effect, attenuation loss, absorption loss, spreading loss and total noise due to thermal Noise, turbulence Noise, shipping Noise, wave Noise. The acoustic channel model used for OFDM Communication will have the noise effect due to attenuation loss and ambient noise (Total Noise). In this research work, In the underwater channel, Multipath parameters such as length, width and depth and attenuation parameters such as frequency, salinity, radial range, SPS and ambient noise parameters such as shipping factor, wave factor are considered in calculating total noise.
Mamatha N, Rakshitha B.S, Chandra Shekar N, Chetan H, Suganya S, and Jyoti M.R
Institute of Advanced Scientific Research
M. R Jyoti, H Chetan, and Manjudevi
American Scientific Publishers