"Vishwa"-International Internship Programme

"Vishwa"-International Internship Program

International Internship Opportunity for Engineering, Architecture, Design and Management Students

The ‘Vishwaniketan-CGC network’ has created opportunities of summer Internships, Master and Ph. D programes for Indian students and teachers, in universities abroad. In last seven years, the network created 25+ tie-ups with universities abroad, more than 900+ students have successfully completed PBL-summer-internships (UG Fellowship), and more than 70 teachers completed Ph.D.

For Engineering, Design & Management

For Architecture

Duration- Engineering/Design/Management : 15th June to 30th

Architecture -15TH May to 15th

"Vishwa"-International Internship Program

International Collaborations For UG Fellowship, MS / Ph.D. Program​

From Europe to America & Southeast Asia, Vishwaniketan’s Global Collaboration set a benchmark for Innovative learning methods & with its International Internship Program, Students get global exposure & training!

0 +

Students completed VIIP Successfully

0 +

Universities in Collaboration

0 +

Projects completed

0 +

Teachers Completed PhD

A brief introduction about 'VIIP'

” A Life transforming Experience”

Apply for VIIP 2022

A Great Opportunity for Students currently enrolled in Engineering, Architecture, Design and Management

Click on the Button below to Apply for VIIP 2022

0

Teesside University

0

Tor Vergata

0

Poznan University

0

Hellnic American University

Register for "Vishwa"-International Internship Program (VIIP)

Projects by Year (Also for Current Year)

Projects by Students

Deep Learning on
Mobile Devices

Project by Maithili Deshmukh, Ashi Parihar & Meeta Hebli

AI in Gaming
& Simulation

Project by Chetan Pathade, Sonia Reddy, Prachit Patil & Shubham Shinde

Optimisation of wind turbine location in wind farm

Project by Nidhi Balaji, Tejas Mahajani & Devyank Patil

Implemention of a Blockchain-based Application for Supply Chain

Project by Swapnil Shinde, Rohan Tanwar & Deepak Khamkar

Social Network analysis by Creating an Alexa skill

Project by Omkar Bhosale, Vaibhav Thakare & Shekhar Tarare

Precast
Panels

Project by Abhijit Waghmare, Rahul Patil & Pranit Patil

Physical Layer Design (Labrotory View)

Project by Gaurav Mhatre, Akash Gupta, Mayank Aggarwal & Vishesh Bhagat

Face Detection using Foreground Segmentation

Project by Rohan Tanwar
& Kavya Rajiv

Fibre Glass
Reinforcement

Project by Vijay Nimbalkar, Siddhesh Gaikwad & Vishwadh Rane

Project List

 

Sr.No.Name of Projects
1“Real-Time V2V communication with a Machine Learning-based System for Detecting Drowsiness of Drivers.”
2"Emotion detection using machine learning"
3"Document classification and extracting key fields from handwritten documents: A machine learning approach"
4

"Application of Internet of Things to Smart Homes, Smart Transportation, and Smart Cities"
Mentors are:

  • Dr. Seshadri Mohan
  • Dr. Mariofanna Milanova
  • Dr. John Talburt.
 

Ural Federal University, Russia

Program Outline

This intensive Summer course is developed in one of the most relevant areas of study. It will give students an insight into such important issues as the quality control of biotechnological products and the processes and methods in food biotechnology.

Short Intern projects 2019
Dates: 18 June 2019 - 28 July 2019

Modules:

Module 1: Quality control of biotechnological products
• Types of quality control and quality characteristics for different biotechnological products.
• Control of organoleptic and basic physico-chemical characteristics of biotechnological products;
• Methods for determination of essential nutrients and macro- and microelements in food stuff
• Xenobiotics of chemical and biological origin in foodstuff and methods of their determination;
• Quality control of specific biotechnological products (yeasts, starters, beer, wine, bread, dairy products);
• Final Test.

Module 2: Processes and Methods in Food Biotechnology
• Areas of food biotechnological manufacturing (Bakery, confectionery, fermenting, yeast, alcohol, dairy, meat and fish processing industries. Production of food additives);
• Types of biotechnological processes and products; Metabolic engineering processes;
• Biocatalysis in Food Biotechnology;
• Alcoholic Beverages Manufacturing (Beer, Wine, Mead, Sider);
• Plant Raw Materials Biotechnological Processing and the Methods of Final Desired Products Characterization;
• Final Test.

And we have one more project for Robotics. We will be able to register students there as well, as it is very new and prospective one:

Project name: Modelling sphero robot using the fundamental principles of mechanics and control.
• A computer based simulation of the operation of a sphero robot using mathematical modelling techniques and applied mathematics and informatics. Students will learn the program and learn he laws of mechanics and mathematical methods that underpin the fundamentals control theory. In the end, students will have a computer based mathematical model of the sphero robot, the program to control it and and animated simulation film. The proposed model is unique and it is not available in any public sources. Students will get acquainted with the existing ‘Sphero’ robots and will gain the skills necessary for their modelling.

Director,
Department of International Educational Programs
Ural Federal University
tel. +7(343) 375-46-89.

Poznan University of Technology

 

Project List 2020

Topics9th of May – 21st of June13th of June – 26th of July
Analysis and modeling of the cycle life of the lithium-ion cells++
Selected issues of electric drive control with PMSM motor++
Tools and techniques in the modern concept of production+ 

Tor Vergata University of Rome

Project list for the students in CS and IT and telecommunications

Project 1:
Experimental evaluation of a link based on LoRa standard (for a team of 3 students, background in telecommunications).
• Report on the stardard LoRa.
• Proposal of an original application of LoRa.
• Experimental evaluation of the link.

Project 2:
Migration with MININET (for a team of 3 students, background in telecommunications, more focused on networking issues).
MININET is a software that enables the emulation of a network. In this project we are going to use it to implement some novel concepts related to SDN (software defined networking).

Project 3:
Virtual network function using Docker (for a team of 3 students, background in Computer Science/telecommunications, key requirement: knowledge of Python).

Project 4:
Design and implementation of a head phantom to perform experiments on radio frequency progation and stroke detection (for a team of 3 students, basic knowledge on propagation of electromagnetics waves, this project could be also adapted for biotechnology background as they must work with chemical substances to implement e realistic phantom).

Project 5:
Analyzing the Effects of Tires Behaviour on the Performance of a Car

Project 6:
Anaerobic Digestion of Organic Wastes and Biomass

Project 7:
Development of a Microgrid Central Controller (MGCC)

Project 8:
Computational Fluid Dynamics for Internal Combustion Engines

Project 9:
Model Predictive Control for Residential Hybrid Energy Systems

Technical University of Sofia

 

Internship topics

India – Bulgaria 2020



Faculty of telecommunications

The Faculty of Telecommunications is the leading educational institution in the field of communication technologies in Bulgaria. It offers advanced Bachelor, Master and PhD Degrees training in "Telecommunications". Telecommunications is one of the most modern and dynamic areas of engineering and technology nowadays. Area, where a huge investment is allocated and many leading companies in the world are working, with the lowest unemployment and a constant need for qualified specialists.

Proposed Topics:
Automatic Discrimination of Speech and Music in Radio Broadcasts
Supervisor: Assoc. Prof. Dr. Ivo Draganov,
E-mail: [email protected]

The discrimination of speech from music is based on the ability to identify any portion of pre- recorded or live audio as containing either type of content. It may find lots of applications, such as recording only of musical content (songs), preserving news podcasts, extraction of advertisement segments and much more for archival, re-broadcasting, or market analysis purposes. Due to the continuous nature of broadcasts and extremely large number of radio stations worldwide it is obligatory the process to be automated without the need of operator intervention. Various methods exist for audio content classification based on feature selection and analysis – zero-crossing rate, average energy of the frame, MEL coefficient analysis and others. Implementing successful application for discrimination of music from speech whether for mobile devices (smartphones, tablets, etc.), desktop computers or dedicated media servers relies on proper combination of extracted features and metrics for measuring the similarity between them. The aim of this project is to test, analyze and select the most efficient similarity measures and the metrics associated with them for automated classification of radio content. Based on the results obtained a sample application should be implemented which demonstrates the operability of the approach. It is recommended, although not obligatory, the prior knowledge of at least one programming language, some experience in signal (image) processing and good knowledge in math (undergraduate level, engineering oriented). The programming environment for testing is Matlab.

Detection and Tracking of Moving Objects in Video
Supervisor: Assoc. Prof. Dr. Ivo Draganov,
E-mail: [email protected]

Detection and tracking of moving objects in pre-recorded or live video streams are important stages in numerous applications for both professional and personal use. They could be used for public roads traffic analysis and detection of critical events (e.g. car crashes), security surveillance of restricted areas (spotting intruders), large number of military applications (detecting, tracking and aiming at various targets), personal use in games based on 3D virtual reality and many more. In order to detect a particular object in a visual scene (captured by a video camera) proper descriptors need to be estimated from separate frames. They include intensity levels change, finding edges and corners with shifted positions from frame to frame, high-order statistical moments (e.g. forming Gaussian Mixture Models, etc.) and others. Once selected for a given application these features need to be combined so the phase of detecting the object of interest could take place. The same or additional features may be used afterwards in order to track the object in time. Its trajectory within the frame or specific isolated locations (e.g. passing through restricted areas, etc.) may be registered and further actions undertaken (e.g. raising an alarm, blocking certain exits, etc.). The aim of the project is to investigate some of the most popular features for object detection and tracking from video by testing them over various video samples. The efficiency of analyzed descriptors in terms of precision and needed processing time should be estimated and then proposition of the most efficient for highway traffic analysis need to be incorporated in demo application. It is recommended, although not obligatory, the prior knowledge of at least one programming language, some experience in signal (image) processing and good knowledge in math (undergraduate level, engineering oriented). The working environment for testing is Matlab.

Object Shape Evaluation and Dimensions Estimation from Images
Supervisor: Assoc. Prof. Dr. Ivo Draganov,
E-mail: [email protected]

Industrial applications relating to manufacturing of large number of details rely on precise estimation of their dimensions and overall shape evaluation in order to comply with common standards and to ensure high reliability of the final products. Non-automated measuring of dimensions is proved to be slow and expensive. Digital imaging offers a way with the use of data processing algorithms to automate the whole process at low price rendering it extremely efficient. In this project various operators applied over digital images of different objects in size and shape will be investigated, e.g. edge and corner detectors, morphology operators, segmentation functions and others. Sequence of processing steps need to be selected for a given type of objects captured at different positions over non-stationary background in order to complete the measurements. Based on them, final decision should be made about the compliance of the manufactured object. Testing with different real-world images will provide wider capabilities for the students to use freely and in a flexible manner the most popular techniques in the field. It is recommended, although not obligatory, the prior knowledge of at least one programming language, some experience in signal (image) processing and good knowledge in math (undergraduate level, engineering oriented). The working environment for testing is Matlab.

Methods and algorithms for automatic 3D object model construction from multiple views
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

Objects which occupy space in the virtual environment can be entities that the user can observe and/or manipulate. So creating consistent and useful models of objects and background of the constrained space is essential. This task explains how to apply mathematical transforms that translate them in the virtual world. This involves two components: Translation (changing position) and rotation (changing orientation). The main goal of this task is to find the best ways to express and manipulate 3D rotations, which are the most complicated part of moving models. Accurate models of already existing complex shaped objects are required for synthesizing arbitrary views and also for recognizing them. Automatic construction of geometric models of 3D objects involves three major steps: (i) data acquisition, (ii) registration of different views, and (iii) integration. Data acquisition involves obtaining either intensity or depth data of an object from multiple viewpoints. Accurate 3D spatial relations between different views may not be easily and directly obtained in many cases. Therefore, integration of data from multiple views is not only dependent on the representation chosen for the model description, but also requires a knowledge of the transformations relating the data obtained from multiple views. The goal of registration, is to find the transformations that relate multiple views, thus bringing the object regions that are shared between them into alignment. Integration merges data from multiple views using the computed view transformations, to create a single surface representation in a unique coordinate frame.

Methods and algorithms for skeleton generation and tracking of skeleton joints for human activity recognition
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

The 3D avatar generates large amount of data points, which have to be sent through the network in real-time. To create the skeleton, a human body is described by a number of joints representing key body parts such as head, neck, shoulders, elbows, wrist, torso, hip, knee and ankles. Each joint is represented by its 3D coordinates. The tracking involves determining all coordinates of these joints in real time to allow fluent interactivity. Multiple sensors are necessary to avoid self-occlusion, which is a common problem among most vision-based sensing systems. However combining the measurements from the different sensors creates a new issue known as the data fusion problem. Based on the captured skeletal data an avatar can be animated. The created 3D model needs to be rigged with the captured skeleton hierarchy and appropriate texture maps. A skeleton based animation strategy must is employed for robustly and accurately fitting the avatar to the skeleton and then large scale deformations and movements can are applied in real-time.

Real-time facial identification, facial features detection and tracking in multi-view environment
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

We will develop and implement methods and algorithms to efficiently identify human faces, including: improved method with increased accuracy for face segmentation; method and algorithm for extracting facial features, after transformation into subspaces for dimensionality reduction and a classifier based on a deep learning neural networks. The method for face segmentation of individuals will be based on sequential combination of the known method of Viola-Jones and convolutional neural network CNN. The developed algorithm will serve to recognize expressions based on facial characteristics. Software development and simulation algorithm to extract facial features will be based on reducing the dimensionality of data by face segmentation in the wavelet space and principal component analysis (PCA). The classification will be made by classifiers such as convolutional neural networks or other types of deep learning neural networks suitable for this application.

Real-time photorealistic animation of the avatar’s body and head movement
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

The main idea of the project is to combine geometry and texture based techniques to animate a personalized avatar. The user’s performance is captured by an RGB-D camera and transferred to the avatar in real-time. We rely on a skeleton based animation to transfer large scale deformations of the body, e.g. walking, jumping or moving the arms. Each joint of the skeleton is represented by its 3D coordinates. The tracking involves determining all coordinates of these joints in real time to allow fluent interactivity. Multiple sensors are necessary to avoid self-occlusion, which is a common problem among most vision-based sensing systems. However combining the measurements from the different sensors creates a new issue known as the data fusion problem so the sensors can work together to correct any inaccurately captured joint data. Based on the captured skeletal data an avatar can be animated. The created 3D model needs to be rigged with the captured skeleton hierarchy and appropriate texture maps. A skeleton based animation strategy will be employed for robustly and accurately fitting the avatar to the skeleton and then large scale deformations and movements will applied in real-time.

Designing and simulating realistic clothing
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

Dressing virtual avatars and animating them with high quality, visually plausible, results is a challenging task. Highly realistic physical simulation of clothing on human bodies in motion is complex: clothing models are laborious to construct, patterns must be graded so that they can be sized to different characters, and the physical parameters of the cloth must be known. Current methods for 3D garment capture are not sufficiently accurate or detailed to compete with physical simulation. Existing capture methods suffer from low resolution, static shapes, simple body motions, capture only one clothing piece, or do not segment the clothing from the body. The main goal of this task is to develop a data-driven clothing capture approach; to capture dynamic clothing on humans from multi-view scans and transform it to more easily dress the virtual avatars. The first step will be capturing the garment geometry in motion on a body, estimate the body shape and pose under clothing, and segment and extract the clothing pieces. Then the captured clothing can be transferred to new body shapes and poses.

Create a 3d model of real object
Supervisor: Assist. Prof. Dr. Nicole Christoff,
E-mail: [email protected]

The aim of this project is to create a parametric modeling of objects and to create a photorealistic avatar. To achieve this, a 3D scanner and / or RGB-D sensors will be used to acquire the necessary data to be processed (point cloud). Before the object size can be calculated, the point cloud must be filtered to segment the object from the surrounding environment. This should be done using an automatic segmentation techniques. After the correct segmentation of the object, an algorithm has to be developed and applied to animate this object.
The deliverable is a report including:
•state-of-the-art on various approaches for modelling of an object;
•improved algoritham for object modelling and slection in 3D space
•create a database of deformable object models
•results from simulation experiments related to the development algoritham
Programming language: C ++ / Java / Matlab
Tools at disposition: 3D scanner Sense and test field table

Scan and avatar yourself
Supervisor: Assist. Prof. Dr. Nicole Christoff,
E-mail: [email protected]

The objective of building human representations is to extract compact, features to encode and characterize a human’s attributes from human shape, pose, and motion, when developing human-centered reasoning systems. Skeleton-based human representations are attractive, due to their robustness to variations of viewpoint, human body scale and motion speed as well as the real-time, online performance. 3D skeleton-based representations are able to model the relationship of human joints and encode the whole body configuration. They are also robust to scale and illumination changes, and can be invariant to camera view as well as human body rotation and motion speed. In addition, many skeleton-based representations can be computed at a high frame rate, which can significantly facilitate online, real-time applications.
The deliverable is a report including:
•state-of-the-art on various approaches for modelling of human body
•improved algoritham for human body modelling with real time application in 3D space taking into account changes in the multi-view environment;
•create a database of deformable human body models
•results from simulation experiments related to the development algoritham
Programming language: C ++ / Java / Matlab
Tool at disposition: 3D scanner or Kinect
Android based speech recognition applications
Supervisor: Prof. Dr. Snejana Pleshkova,
E-mail: [email protected]

User can control a variety of applications on an android based platform, which include native applications as well as user installed applications with voice commands. These include - calling, texting, switching on and off sensors (Wi-Fi, GPS, Bluetooth),setting alarms. The application provides online as well as offline services. The application also applies machine learning concepts to identify usage patterns and create an environment which anticipates user requirements. The tasks being performed repetitively are automated. Services of activity recognition, recognizing nearby friends using Bluetooth are performed. The importance of the project is that it provides visually challenged people as well as the general population an alternate and a very easy way to control applications on android smart phones.

Wireless and Mobile Audio Streaming Technologies
In this project the students will learn the principles and structures of wireless and mobile audio streaming technologies, the architecture and block diagrams of wireless and mobile audio systems. They will design the basic algorithms for audio streaming via wireless and mobile audio professional or home networks developing real working applications for audio information streaming and also applications of remote control of wireless and mobile audio system using smart phones or tablets.
PREREQUISITES: Audio Systems, Wireless and Mobile Networks, Audio Streaming, Matlab, Simulink, Android Studio, Eclipse IDE, Web Design

Microphones, Microphone Arrays Based on MEMS Technology and Applications in Audio Visual Mobile Robots Motion Control
In this project the students will learn the MEMS microphones technology. Using the fundamental theory of microphone arrays they will develop practical applications of different structures of microphone arrays, testing them and analysing their ability to determine the direction of sound of arrival from speakers or other audio sources and using the results in real time audio visual mobile roots motion control.
PREREQUISITES: Audio Systems, Microphones, Microphone Arrays, MEMS Microphone Technology, Microcontrollers Hardware and Software, Microsoft Mobile Robots Studio

Room Acoustic Analysis. Simulations an Real Implementations
In this project the students will acquire basic knowledge and skills about principles of acoustic, specific and principles of room acoustic for audio studios, concert halls, offices, home rooms, etc. These knowledge’s will be applied from the students in simulations and practical implementations of real room acoustic analysis, taken the results and conclusions for choosing and using appropriates building materials and building constructions to greatly improve the acoustic characteristics of the room under test.
PREREQUISITES: Audio Systems, Acoustic, Room Acoustic, Audio Signal Processing, Neural Networks, Matlab, Simulink, Microsoft Visual Studio, Java, Sound Insulation Materials and Constructions

Audio Systems for Creation, Editing and Mastering of Songs and Musical Productions
In this project the students will learn the principles, the architectures and basic functions of audio systems for creation, editing, processing and mastering of songs or other musical production. Then students will work with the professional audio editing system PreSonus Studio to develop real practical applications producing the new songs, applying all necessary steps for creation of audio and instrumental tracks, adding the special audio effects, editing and processing all created tracks to form the final release of the created song.
PREREQUISITES: Audio Systems, Editing Audio Systems, Audio Signal Processing, Special Audio Effects, Audio Mastering.

Faculty of Mechanical Engineering

Supervisor: Assist. Prof. Dr. Kalin Chuchuganov,
E-mail: [email protected]

Condition monitoring and analysis of ball bearings, using vibration monitoring software – The condition of ball bearings is analyzed using specially developed diagnostic stand and by taking vibration spectra, which are analyzed in vibration monitoring software in Labview environment.
Dynamic balance analysis and methods for balancing of rotating machinery – vibration spectra of rotating machinery is examined in order to assess their balance condition using vibration analysis hardware and software in Labview environment. Consequently methods for balancing are used and a secondary assessment of the balance condition is performed to evaluate the balancing process. The test object used for research is a computer fan.
Analysis of the quality of color printing of laser and LED color printers, used for prepress hard-proofing and digital offset –the conformance to printing standards for offset printing and hard-proofing of digital color printers is assessed qualitatively and quantitatively using spectrophotometers, calibration and profiling and software packages for printer quality evaluation and color management. The research is made in a specially developed color management and research laboratory.
Analysis of color reproduction abilities of color LCD displays, used in prepress and digital color laser printing – color LCD displays are evaluated with the use of color management software, profiling and calibration procedures with spectrophotometer. The conformity of the displays with the standards for digital color printing and prepress is assessed. The research is made in a specially developed color management and research laboratory.
Quality evaluation of illumination lighting systems, used in premises for prepress and finished print production examination –the spectra and other parameters of the light, emitted from lighting systems, used in the printing industry are assessed for conformity with the specifications of the standards in the field. The research is made in a specially developed color management and research laboratory.

Faculty of Electronic Engineering and Technologies

Faculty of Electronic Engineering and Technologies (FEET) is recognized as a leader in education and research in the field of electronics at national level (accreditation by Institution of Electrical Engineering, London, UK). Training in FEET is consistent with the latest advances in electronics, as well as with educational and research programs of leading European universities in England, Germany, France, The Netherlands, Italy, etc.

Supervisor: Assoc. Prof. Dr. Petar Yakimov,
E-mail: [email protected]

Subjects: Students will have the opportunity to learn about the principles of operation and to explore various transformer devices used as mains power supplies. Research is conducted on real-power models (rectifiers, stabilizers, UPS, photovoltaic systems, rechargeable batteries, etc.). Measuring points are shown on the test devices to illustrate the principles of operation and the basic electrical ratios in these circuits. An opportunity for computer simulation of individual schemes has been created. Through simulation studies with specialized software, it is possible to compare the results with real devices.

 

 

Internship topics

India – Bulgaria 2019



Faculty of telecommunications

The Faculty of Telecommunications is the leading educational institution in the field of communication technologies in Bulgaria. It offers advanced Bachelor, Master and PhD Degrees training in "Telecommunications". Telecommunications is one of the most modern and dynamic areas of engineering and technology nowadays. Area, where a huge investment is allocated and many leading companies in the world are working, with the lowest unemployment and a constant need for qualified specialists.

Proposed Topics:
Automatic Discrimination of Speech and Music in Radio Broadcasts
Supervisor: Assoc. Prof. Dr. Ivo Draganov,
E-mail: [email protected]

The discrimination of speech from music is based on the ability to identify any portion of pre- recorded or live audio as containing either type of content. It may find lots of applications, such as recording only of musical content (songs), preserving news podcasts, extraction of advertisement segments and much more for archival, re-broadcasting, or market analysis purposes. Due to the continuous nature of broadcasts and extremely large number of radio stations worldwide it is obligatory the process to be automated without the need of operator intervention. Various methods exist for audio content classification based on feature selection and analysis – zero-crossing rate, average energy of the frame, MEL coefficient analysis and others. Implementing successful application for discrimination of music from speech whether for mobile devices (smartphones, tablets, etc.), desktop computers or dedicated media servers relies on proper combination of extracted features and metrics for measuring the similarity between them. The aim of this project is to test, analyze and select the most efficient similarity measures and the metrics associated with them for automated classification of radio content. Based on the results obtained a sample application should be implemented which demonstrates the operability of the approach. It is recommended, although not obligatory, the prior knowledge of at least one programming language, some experience in signal (image) processing and good knowledge in math (undergraduate level, engineering oriented). The programming environment for testing is Matlab.

Detection and Tracking of Moving Objects in Video
Supervisor: Assoc. Prof. Dr. Ivo Draganov,
E-mail: [email protected]

Detection and tracking of moving objects in pre-recorded or live video streams are important stages in numerous applications for both professional and personal use. They could be used for public roads traffic analysis and detection of critical events (e.g. car crashes), security surveillance of restricted areas (spotting intruders), large number of military applications (detecting, tracking and aiming at various targets), personal use in games based on 3D virtual reality and many more. In order to detect a particular object in a visual scene (captured by a video camera) proper descriptors need to be estimated from separate frames. They include intensity levels change, finding edges and corners with shifted positions from frame to frame, high-order statistical moments (e.g. forming Gaussian Mixture Models, etc.) and others. Once selected for a given application these features need to be combined so the phase of detecting the object of interest could take place. The same or additional features may be used afterwards in order to track the object in time. Its trajectory within the frame or specific isolated locations (e.g. passing through restricted areas, etc.) may be registered and further actions undertaken (e.g. raising an alarm, blocking certain exits, etc.). The aim of the project is to investigate some of the most popular features for object detection and tracking from video by testing them over various video samples. The efficiency of analyzed descriptors in terms of precision and needed processing time should be estimated and then proposition of the most efficient for highway traffic analysis need to be incorporated in demo application. It is recommended, although not obligatory, the prior knowledge of at least one programming language, some experience in signal (image) processing and good knowledge in math (undergraduate level, engineering oriented). The working environment for testing is Matlab.

Object Shape Evaluation and Dimensions Estimation from Images
Supervisor: Assoc. Prof. Dr. Ivo Draganov,
E-mail: [email protected]

Industrial applications relating to manufacturing of large number of details rely on precise estimation of their dimensions and overall shape evaluation in order to comply with common standards and to ensure high reliability of the final products. Non-automated measuring of dimensions is proved to be slow and expensive. Digital imaging offers a way with the use of data processing algorithms to automate the whole process at low price rendering it extremely efficient. In this project various operators applied over digital images of different objects in size and shape will be investigated, e.g. edge and corner detectors, morphology operators, segmentation functions and others. Sequence of processing steps need to be selected for a given type of objects captured at different positions over non-stationary background in order to complete the measurements. Based on them, final decision should be made about the compliance of the manufactured object. Testing with different real-world images will provide wider capabilities for the students to use freely and in a flexible manner the most popular techniques in the field. It is recommended, although not obligatory, the prior knowledge of at least one programming language, some experience in signal (image) processing and good knowledge in math (undergraduate level, engineering oriented). The working environment for testing is Matlab.

Methods and algorithms for automatic 3D object model construction from multiple views
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

Objects which occupy space in the virtual environment can be entities that the user can observe and/or manipulate. So creating consistent and useful models of objects and background of the constrained space is essential. This task explains how to apply mathematical transforms that translate them in the virtual world. This involves two components: Translation (changing position) and rotation (changing orientation). The main goal of this task is to find the best ways to express and manipulate 3D rotations, which are the most complicated part of moving models. Accurate models of already existing complex shaped objects are required for synthesizing arbitrary views and also for recognizing them. Automatic construction of geometric models of 3D objects involves three major steps: (i) data acquisition, (ii) registration of different views, and (iii) integration. Data acquisition involves obtaining either intensity or depth data of an object from multiple viewpoints. Accurate 3D spatial relations between different views may not be easily and directly obtained in many cases. Therefore, integration of data from multiple views is not only dependent on the representation chosen for the model description, but also requires a knowledge of the transformations relating the data obtained from multiple views. The goal of registration, is to find the transformations that relate multiple views, thus bringing the object regions that are shared between them into alignment. Integration merges data from multiple views using the computed view transformations, to create a single surface representation in a unique coordinate frame.

Methods and algorithms for skeleton generation and tracking of skeleton joints for human activity recognition
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

The 3D avatar generates large amount of data points, which have to be sent through the network in real-time. To create the skeleton, a human body is described by a number of joints representing key body parts such as head, neck, shoulders, elbows, wrist, torso, hip, knee and ankles. Each joint is represented by its 3D coordinates. The tracking involves determining all coordinates of these joints in real time to allow fluent interactivity. Multiple sensors are necessary to avoid self-occlusion, which is a common problem among most vision-based sensing systems. However combining the measurements from the different sensors creates a new issue known as the data fusion problem. Based on the captured skeletal data an avatar can be animated. The created 3D model needs to be rigged with the captured skeleton hierarchy and appropriate texture maps. A skeleton based animation strategy must is employed for robustly and accurately fitting the avatar to the skeleton and then large scale deformations and movements can are applied in real-time.

Real-time facial identification, facial features detection and tracking in multi-view environment
Supervisor: Assoc. Prof. Dr. Agata Manolova,
E-mail: [email protected]

We will develop and implement methods and algorithms to efficiently identify human faces, including: improved method with increased accuracy for face segmentation; method and algorithm for extracting facial features, after transformation into subspaces for dimensionality reduction and a classifier based on a deep learning neural networks. The method for face segmentation of individuals will be based on sequential combination of the known method of Viola-Jones and convolutional neural network CNN. The developed algorithm will serve to recognize expressions based on facial characteristics. Software development and simulation algorithm to extract facial features will be based on reducing the dimensionality of data by face segmentation in the wavelet space and principal component analysis (PCA). The classification will be made by classifiers such as convolutional neural networks or other types of deep learning neural networks suitable for this application.

Faculty of Mechanical Engineering

Supervisor: Assist. Prof. Dr. Kalin Chuchuganov,
E-mail: [email protected]

Condition monitoring and analysis of ball bearings, using vibration monitoring software – The condition of ball bearings is analyzed using specially developed diagnostic stand and by taking vibration spectra, which are analyzed in vibration monitoring software in Labview environment.
Dynamic balance analysis and methods for balancing of rotating machinery – vibration spectra of rotating machinery is examined in order to assess their balance condition using vibration analysis hardware and software in Labview environment. Consequently methods for balancing are used and a secondary assessment of the balance condition is performed to evaluate the balancing process. The test object used for research is a computer fan.
Analysis of the quality of color printing of laser and LED color printers, used for prepress hard-proofing and digital offset – the conformance to printing standards for offset printing and hard-proofing of digital color printers is assessed qualitatively and quantitatively using spectrophotometers, calibration and profiling and software packages for printer quality evaluation and color management. The research is made in a specially developed color management and research laboratory.
Analysis of color reproduction abilities of color LCD displays, used in prepress and digital color laser printing – color LCD displays are evaluated with the use of color management software, profiling and calibration procedures with spectrophotometer. The conformity of the displays with the standards for digital color printing and prepress is assessed. The research is made in a specially developed color management and research laboratory.
Quality evaluation of illumination lighting systems, used in premises for prepress and finished print production examination – the spectra and other parameters of the light, emitted from lighting systems, used in the printing industry are assessed for conformity with the specifications of the standards in the field. The research is made in a specially developed color management and research laboratory.

Faculty of Electronic Engineering and Technologies

Faculty of Electronic Engineering and Technologies (FEET) is recognized as a leader in education and research in the field of electronics at national level (accreditation by Institution of Electrical Engineering, London, UK). Training in FEET is consistent with the latest advances in electronics, as well as with educational and research programs of leading European universities in England, Germany, France, The Netherlands, Italy, etc.

Supervisor: Assoc. Prof. Dr. Petar Yakimov,
E-mail: [email protected]

Subjects: Students will have the opportunity to learn about the principles of operation and to explore various transformer devices used as mains power supplies. Research is conducted on real-power models (rectifiers, stabilizers, UPS, photovoltaic systems, rechargeable batteries, etc.). Measuring points are shown on the test devices to illustrate the principles of operation and the basic electrical ratios in these circuits. An opportunity for computer simulation of individual schemes has been created. Through simulation studies with specialized software, it is possible to compare the results with real devices.

Athens Information Technology, Greece

Summer 2019 UG Fellowship Programme

 

Project Descriptions


Project Number: 1
Project Title: Video Analytics: Face normalization
AIT supervisors’ name(s) and URLs:
Aristodemos Pnevmatikakis and Stefanos Astaras
https://www.ait.gr/team/aristodemos-pnevmatikakis/
https://www.ait.gr/team/stefanos-astaras/
Estimated group size: 3
Estimated duration (3 weeks/6 weeks): 3 weeks
Project description: Face normalization: Simple normalization: Affine or projective transform, Face morphing for normalization

Project Number: 2
Project Title: Video Analytics: Recognition
AIT supervisors’ name(s) and URLs:
Aristodemos Pnevmatikakis and Stefanos Astaras
https://www.ait.gr/team/aristodemos-pnevmatikakis/
https://www.ait.gr/team/aristodemos-pnevmatikakis/
Estimated group size: 3
Estimated duration (3 weeks/6 weeks): 3 weeks
Project description:Face, Age, Gender, Emotion recognition: Train classifier Bayesian, HOG, CNN), use to recognize

Project Number: 3
Project Title: Video Analytics: Context in images
AIT supervisors’ name(s) and URLs:
Aristodemos Pnevmatikakis and Stefanos Astaras
https://www.ait.gr/team/aristodemos-pnevmatikakis/
https://www.ait.gr/team/stefanos-astaras/
Estimated group size: 3
Estimated duration (3 weeks/6 weeks): 3 weeks
Project description:Context in images: Experiment with a deep learning CNN that recognizes 1000 types of objects. Evaluate performance

Project Number: 4
Project Title: SW development for Android and the web
AIT supervisors’ name(s) and URLs:
Aristodemos Pnevmatikakis and George Lalas
https://www.ait.gr/team/aristodemos-pnevmatikakis/
https://www.ait.gr/team/george-lalas/
Estimated group size: 3
Estimated duration (3 weeks/6 weeks): 6 weeks
Project description: Android GPS logger stores locally the GPS location of the user. A backend collects the stored data, and presents them to the Android clients of users that are accepted as friends. This is a project in two parts, to actually take place for the entire duration of the students' stay @ AIT

Project Number: 5
Project Title: Introduction to text mining on social networks
AIT supervisors’ name(s) and URLs:
Fotis Talantzis
https://www.ait.gr/team/fotios-talantzis/
Estimated group size: 2
Estimated duration (3 weeks/6 weeks): 3 weeks
Project description: Vast amounts of new information and data are generated everyday through economic, academic and social activities, much with significant potential economic and societal value. Techniques such as text and data mining and analytics are required to exploit this potential. Using data originating from major news providers (e.g. BBC, CNN etc.) the student is expected to implement a web-based system that:
• Extracts and Collects most significant keywords from news items and stores them in a database including their category (e.g. Sports, Finance etc.)
• Creates summary from large chunks of text (news items)
• Given a specific news item, suggest similar news items based on the extracted keywords
• Given a specific news item, suggest its originating category (e.g. Sports, Finance etc.)

Project Number: 6
Project Title: Analysis of social profiles
AIT supervisors’ name(s) and URLs:
Fotis Talantzis
https://www.ait.gr/team/fotios-talantzis/
Estimated group size: 2
Estimated duration (3 weeks/6 weeks): 3
Project description: Social network analysis involves studying the relationships between a set of users. In many situations, there are patterns to the types of relationships that are formed - for example, communities of people who are more likely to link to each other than to other people in the network. Using Twitter as an example students will start with a Twitter user and identify relationships of the users that follow him (followers) by recursively examining the followers of each of the followers of the original user. The student is expected to implement a web- based system that
• Interfaces with the Twitter API
• Fetches and prints the followers of a user identified by their username
• Given a depth (integer) fetch the followers of each of the followers of the user
• Generate statistics e.g. Print those users that are followed by more users in the examined network

Project Number: 7
Project Title: Blockchain support for supply chain
AIT supervisors’ name(s) and URLs:
Sofoklis Efremidis
https://www.ait.gr/team/sefremidis/
Estimated group size: 3
Estimated duration (3 weeks/6 weeks): 6
Project description: Blockchain is a relatively new technology that first appeared along with the bitcoin cryptocurrency. Besides being a foundational technology for a number of other cryptocurrencies, the potential of blockchains to support new innovative applications is still explored. In this project a sample application for tracking goods in a supply chain will be developed. The application will use a blockchain for tracking the transfer of goods between stakeholders of the supply chain and will use escrow accounts for implementing payments between them.

Project Number: 8
Project Title: Blockchain support for implementing Service Level Agreements
AIT supervisors’ name(s) and URLs:
Sofoklis Efremidis
https://www.ait.gr/team/sefremidis/
Estimated group size: 3
Estimated duration (3 weeks/6 weeks): 6
Project description: Blockchain is a relatively new technology that first appeared along with the bitcoin cryptocurrency. Besides being a foundational technology for a number of other cryptocurrencies, the potential of blockchains to support new innovative applications is still explored. Smart contracts can be implemented on a blockchain for automatically enforcing business rules and Service Level Agreements. In this project smart contracts will be developed for this purpose. In an environment where Service Level Agreements have to be enforced between federated service providers and service consumers.

Project Number: 9
Project Title: Blockchain support of IoT applications
AIT supervisors’ name(s) and URLs:
Sofoklis Efremidis
https://www.ait.gr/team/sefremidis/
Estimated group size: 3
Estimated duration (3 weeks/6 weeks): 6
Project description: Blockchain is a relatively new technology that first appeared along with the bitcoin cryptocurrency. Besides being a foundational technology for a number of other cryptocurrencies, the potential of blockchains to support new innovative applications is still explored. In this project blockchains will be used for IoT support for anonymized data and metrics collection and credit point redemption for a retail marketing setting.

Project Number: 10
Project Title: Planar Antenna Design
AIT supervisors’ name(s) and URLs:
Dimitrios Ntaikos
https://www.ait.gr/team/dimitriosntaikos/
Estimated group size: 4 students (2 teams, 2 students each)
Estimated duration (3 weeks/6 weeks): 3
Project description: For this Project, students will become familiar with the basic knowledge of antennas and electromagnetic theory. After discussion with the supervisor, each team will be appointed with an antenna design, for which they will perform a detailed study of this specific planar antenna. They will calculate on paper, using the analytical mathematical expressions and equations provided, the physical dimensions and characteristics of the proposed antenna. They will run electromagnetic simulations of the chosen planar antenna. They will write a report summarizing their findings. Requirements: Knowledge of MatLab, Mathematics (Complex and Vector Calculus).

Project Number: 11
Project Title: Sub-6 GHz Wireless Over-the-Air Transmission
AIT supervisors’ name(s) and URLs:
Ioannis Chondroulis
Estimated group size: 4 students (2 teams, 2 students each)
Estimated duration (3 weeks/6 weeks): 3
Project description: This project’s objective is to transmit over-the-air data in frequencies below 6 GHz using the B-WiSE lab’s software- defined radio modules. During the course of the project, wireless communication theory, MATLAB programming and simulation, as well as key aspects of hardware design will be covered. In particular, students will become familiar with the physical layer of the B-Wise lab’s WARP programmable radio modules and have the chance to transmit and receive actual data over the air. A number of experiments involving different antenna configurations and user setups will be performed. A respective report will be composed.
Requirements: Basic network knowledge, familiarity with Matlab.

Project Number: 12
Project Title: Detecting Addictive Behavior in Online Games of Chance
AIT supervisors’ name(s) and URLs:
Ioannis Christou
https://www.ait.gr/team/ioannischristou/
Estimated group size: 1 student
Estimated duration (3 weeks/6 weeks): 3
Project description:
Abstract: Online Games of Chance are increasingly confronted with addictive behavior from certain subscribers of their services; to tackle such problems that affect both the players, as well as the companies themselves, the latter, in order to obtain certain accreditations and certifications, are obliged to show that they have significant Responsible Gaming Initiatives installed. According to the DSM-5 manual for addictive behaviors, two common tell-tale signs of addictive behavior is the continuous gambling for long periods of time despite accumulated losses, and the “doubling of bets” phenomenon where the player more than doubles their bets with each gamble in an effort to break even.

The project involves therefore developing a system for detecting either one of the above two conditions in CSV files containing historical data of anonymous players. The objective is to develop efficient Java code that will parse such files, and detect if some players exhibit such behavior, taking into account some necessary user-defined parameters (such as what constitutes a time-window, and how many bets must a player place in order to raise an alarm).

Pre-requisites: excellent knowledge of Java, Source-code control systems (Git), Data Structures, Data Bases, Algorithms.

Mae Fah Luang University, Thailand

 

Project Based International Summer Internship Program 2019

 

List of Projects Offered for Internship



1.Computer (School of Information Technology)

Sr.No.TopicAdvisorNumber of students
1Personalized Learning(Mobile Learning)Assist. Prof. Dr. Punnarumol Temdee2
2Signal Processing: Frequency estimationAssist. Prof. Dr. Nattapol Aunsri2
3Signal Processing: Transforms techniquesAssist. Prof.Dr. Nattapol Aunsri2
4IoTs Application for AgricultureDr. Chayapol Kamyod2

Student Requirement: Student must have their own Laptop to work on project.



2.Civil (School of Science)

Sr.No.TopicAdvisorNumber of students
1Cement/Geopolymer Composite for Construction
To prepare cement or geopolymer based materials with addition of some specialty materials. Testing such as strength, water absorption, and thermal conductivity.
1. Assoc Prof.Dr. Darunee Wattanasiriwech
2. Assist. Prof. Dr. Suthee Wattanasiriwech
3. Dr. Sithi Duangpet
6
1Engineered Wood Based on Rice Straw
To prepare Bio-composites with different structure and test their properties such as strength and water resistance.
1. Assist. Prof.Dr. Nattakan Soykeabkaew
2. Assist. Prof.Dr. Nattaya Tawichai
3. Assist. Prof.Dr. Uraiwan Tntatha

University of Nevada ,Las Vegas

 

Project List

 

Sr.No.Name of Projects
1The projects are in the area of robotics and automation design, research and development, and testing
21) Developing a predictive model for diabetic ulcers
2) Biomechanics of walking under reduced gravity
3) Shock transmission in bolted joints
4) Design of demand-activated rumble strips
31) Heat exchanger design. Knowledge of using ANSYS and Fluent commercial software is highly desirable
41) Biosensing
2) Nanotechnology
3) Photovoltaics
4)Biomaterials
5The projects are in the area of autonomous sensing and navigation of small scale robotic flying platforms for source localization and contour mapping.

Aarhus University, Denmark.

The students will get:
• An intense Multi Business Model training the first 2-3 weeks.

• CGC certified researchers and coaches will use B-labs and MBIT Lab to train students in our MBIT tools, mindset and theories.

• They will follow about the same curriculum as we use to AU students.

• Attached are videos as examples of how we work with your students:
https://vimeo.com/269442986
https://vimeo.com/320344850
https://vimeo.com/manage/videos

• CGC – AU researchers will train and present the students to relevant technologies that are related to our chosen TBMI challenges that the students will have to solve and pitch to judges at the end of their stay.

• TBMI Challenges will be cross interdisciplinary and we will therefore form the groups as such. TBMI Challenges will have Danish business as “host” for the TBMI Challenge.

• We expect TBMI Challenges within drone, water, knowledge home and CGC researchers will present and work with students on technical and engineering theory relevant to the challenges.

• Your student will work both with physical and digital training tools developed by CGC.

• We need students that are open minded, has a cross interdisciplinary approach/mindset and are international/Global oriented. We need students that wants to “bridge technology and business”.

• We aim at training the students so they can bring back the mindset, knowledge and Competences of The Multi Business Model Innovation and Technology CGC approach to India and hopefully build up many B-Labs/B-cubes in India – and globally.