Dr. T. V. Smitha
Brief Description of the VGST Project on Video Processing Application
Video processing applications often need to incorporate robust noise reduction techniques that provide visually pleasing effects. For real-time implementation, a novel hardware architecture has been developed which enables noise reduction and object tracking. This project implements automated subparametric mesh-based optical flow estimation and video quality enhancement using Convolutional Neural Networks (CNN). In hardware implementation mesh based median filter with the Lucas-Kanade approach has been used for noise removal and object tracking.
A versatile curved triangle mesh generator for finite element-based image/video processing applications using MATLAB. This approach first creates meshes using an open-source finite element mesh generator. The proposed technique then generates and visualises cubic-order curved triangular elements using subparametric transformation. The resulting meshes can be used to solve a variety of problems using finite element analysis or mesh-based techniques for image and video processing applications. This technology can be used in medical simulation, image processing, material engineering, video processing, computational materials science, optical flow estimation, computational electromagnetics, etc.
Further, a general framework to create an accurate algorithm for video enhancement tasks, including deblurring and denoising, is developed.
An automated subparametric curved triangular mesh generator for video frames in Python from Gmsh as part of the development of the proposed framework is created. At each nodal point, the optical flow values are then calculated. Using the TensorFlow and autoencoder technique, denoising is applied to each video frame. Thus, each frame’s video quality is improved by using optical flows and CNN algorithms. Hence, a general architecture that integrates the flows with various task-specific layers, building on top of the flow estimation, is developed.
Results of Simulation:
Results of video quality enhancement algorithm developed using CNN based autoencoder and optical flow estimation from the subparametric cubic-curved meshed video frames:
Input Images of the Testing dataset:
FEM
Mr. Chandan K
The research interests revolve around leveraging the power of finite element method (FEM)-based simulations to investigate and tackle complex engineering and scientific problems, and focus on areas such as heat transfer and fluid dynamics. The research is supported by utilizing software such as ANSYS, COMSOL, and FEniCS, and their implementation enables to model and simulate intricate systems. Consider varying material properties and boundary conditions to gain valuable insights into the behaviour and performance of structures and phenomena. The aim is to contribute to advancements in fields such as structural engineering, mechanical engineering, and computational physics, ultimately leading to improved designs, optimized processes, and enhanced understanding of physical phenomena.
Machine Learning/DL/AI
The field of machine learning (ML), deep learning (DL), and artificial intelligence (AI) has witnessed a tremendous surge in interest, and its applications have expanded to various domains, including engineering. These techniques offer a powerful framework to tackle complex engineering problems by leveraging vast amounts of data and extracting meaningful insights. The research applied to areas such as predictive maintenance, optimization of engineering systems, anomaly detection, image and signal processing, robotics, and automation are currently trending and integrating them to real world engineering problems is an emerging field. By integrating them into engineering workflows where engineers can enhance decision-making processes, improve system performance, and develop innovative solutions that were previously unattainable. The growing interest in ML/DL/AI reflects the recognition of its potential to revolutionize engineering practices, enabling the development of intelligent systems that can adapt, learn, and optimize their performance in real-world scenarios.
Solid Mechanics
Dr. R. S. Varun Kumar
Solid mechanics is a collection of physical laws, mathematical techniques, and computer algorithms that can be used to predict the behavior of a solid material that is subjected to mechanical or thermal loading. All those who seek to understand natural phenomena involving the stressing, deformation, flow, and fracture of solids, and all those who would have knowledge of such phenomena to improve our living conditions and accomplish human objectives, have use for solid mechanics. Natural phenomena involving solid mechanics are studied in geology, seismology, and tectonophysics, in materials science and the physics of condensed matter, and in parts of biology and physiology. The field has a wide range of applications, including geomechanics, mechanical engineering, biomechanics, materials science, and nanotechnology.
Nanofluid: Artificial neural networks (ANNs) are a significant aspect of the new artificial intelligence sector. They are a model for machine learning (ML) and have become considerably effective compared to traditional regression and statistical approaches. ANN applications are now widely utilized in a variety of human requirements. In a broad sense, an ANN functions as a human brain simulation. The innovative structure of the data processing framework is a crucial aspect of this approach. It has revealed valuable intelligence and has solved real-world issues that modern computers cannot resolve. ANNs are capable of handling problems in agriculture, science, medical science, education, finance, management, security, engineering, trading commodity, and art. Including problems in manufacturing, transportation, computer security, banking, insurance, properties management, marketing, energy, and those challenges that cannot be solved by the computational ability of traditional procedures and conventional mathematics. ANN has the advantages of adaptive, self-organizing, and real-time learning features. It has the capacity to find the best solutions quickly. ANNs have recently become a popular and useful model for classification, clustering, pattern recognition, and prediction in various fields.
Computational Fluid Dynamics
Dr. Madhukesh J. K.
Computational fluid dynamics is an assertive discipline that includes fluid mechanics, numerical analysis, and computer science for the simulation and analysis of fluid flow patterns. Using mathematical modeling and complex algorithm, CFD makes a path for engineers and scientists to study a wide range of phenomena, including motion of fluid motion, heat, and mass transfer in various systems. CFD simulations allow engineers to evaluate the performance, optimization of designs, and identify potential problems. It also helps in understanding complex flow patterns, including boundary layer separation. CFD is predicted to be more significant in addressing complex fluid dynamics issues, contributing to developments in numerous sectors and a better knowledge of fluid behavior.
Nanofluid: Nanofluids, nanoparticle suspensions in a base fluid, have emerged as a potential subject in recent years due to their unique thermal and fluidic characteristics. Adding nanoparticles, such as metallic or non-metallic particles, to the base fluid drastically modifies its behavior, resulting in increased thermal conductivity, faster heat transfer rates, and altered fluid flow properties. This unique class of fluids is essential and has several applications in various industries.
Because of their increased thermal conductivity, nanofluids are helpful in thermal management applications such as cooling electronic devices, heat exchangers, and renewable energy systems. By increasing heat transfer rates, nanofluids can improve energy conversion efficiency and lower operational costs. Furthermore, nanofluids use sophisticated heat transfer fluids, lubricants, and coolant additives, where their improved characteristics can lead to enhanced system performance and lifetime.
The relevance and extent of nanofluids are predicted to grow as research progresses. Nanofluids hold the potential to transform numerous industries and pave the path for innovative applications in the fields of energy, electronics, healthcare, and beyond, with ongoing investigations into the creation of new nanoparticles, optimization of their dispersion, and understanding of their fundamental behavior.
