Hot Spot Prediction and Detection in Grain Storage

This project is with colaboration with TeleSense in Pakistan, that is basically working on grain storage especially corn. They made a set up in Pakistan, that are grain storage silos and put sensor there. They get data of carbon dioxide, weather temperature, and moisture. Sensors are connected with WiFi and are sending data on cloud continuously on hourly basis. Through our supervisor who is in touch with that company, provided us data sheet for further process. Now, here I am describing what is going on in that silo and what we are doing. People present in silo are looking at the data of temperature provided by sensors and calculating manually, whether hot spot will be generated or not. They don’t have a proper automated system for this purpose. They are doing calculations manually that is really a time taking procedure. It also need more people for maintenance. We are using machine learning and deep learning models to automate this system.

Machine Learning AgriTech

Smart Gas Leakage Detector (SGLD)

We have been thriving for a product that could save lives from gas evacuation from house hold gas usages. We have named our product as Smart Gas Leakage Detector. This entire system would be helpful in reducing the effects of gas leakage on the lives of humans and it would also save environment. The entire system is being developed by using different technologies and skills. We as group members worked as a team to collaborate for the formation of a good product.

Internet of Things (IoT) Cloud Computing

Plant disease diagnosis using Machine Learning

Agriculture is the lifeline of Pakistan’s economy. It plays a central role in national development, food security and poverty reduction. The focus is on enhancing productivity, without considering the ecological impacts that has resulted in environmental degradation. As disease of the plants is inevitable, detecting disease plays a major role in the field of agriculture. Potato and Tomato are among the most significant vegetables and cash crops in Pakistan. These crops are highly affected by diseases. Farmers face many difficulties in detecting these diseases. They have to visit to laboratories or hire an export for continuously monitoring. These are lengthy and time consuming processes. They spent money on getting results. These results also take time. In that testing time, there crops go down. Keeping in view that thing, we are providing our farmers with an android application through which they can detect diseases of potato and tomato. After that, they can treat it easily and on time. All he has to do is taking the picture of affected part of crop and upload it to our application. Then our deep learning trained model will classify the image and will predict the disease name and its treatment as a result. Model will be on the server side. We will convolutional neural network as a model as it extracts the feature from image itself. Application interface will be simple and easy to use. As our most of farmers are uneducated. So, application will be available in Urdu language as well.

Machine Learning Mobile Application Development Computer Vision Agri Tech

GoMeet

An android app which connects different local service providers like tailors, carpenters, barbers, cobblers etc to different customers locally. The types of apps would connect local service providers to more customers and would provide them the platform to reach more and more customers.

Mobile Application Development

Intelligent Medical Consultant

The Healthcare system plays a significant role in the progress of a country. In Pakistan, especially in its rural areas, healthcare system is not good enough to facilitate the rural population that is more than 60% of the total population of Pakistan and one reason for this is the unavailability of specialist doctors. Almost half a million people die annually due to errors in diagnostics and medication in Pakistan. By seeing this situation of healthcare system, we aimed to develop an application for doctors that will help them to minimize diagnostics errors, especially in the rural areas of Pakistan.

Machine Learning Information Retrievel

A Moody Deep Fake

I am doing FYP in Deep Fake technology where I am using two pre-trained models including Face2Face and a Mood Generator model entitled First Order Motion Model and GANimation. I will have three inputs one driving video of a face (Source A), and a still face image (Final Target) and finally another still face image (Reference Emotion). The first part consists of using FOMM to transfer facial expressions in the video (Source A) to the face in the Target to get Target image as deep fake video where number of frames will be equal to that of Source A video. The second part is to use GANimation model to covert the emotion of deep fake video based on the emotion of Reference Emotion image and this will give us A Moody Deep Fake, where the deep fake is in a particular mood. To make whole system more user-friendly, I have also developed a Flutter app which can be used in any platform including Mobile, Web and Desktop. The app is connected to python server written in Flask which receives images and video as http request performs operation based on requested model and results are sent back as a response to http request.

Computer Vision Machine Learning

Zombie Apocalypse

Game Development

Stock Market Prediction using Sentiment Analysis

Investment in the stock market is risky, but when approached in a disciplined manner, it is one of the most efficient ways to enjoy substantial profits. Investors evaluate the performance of a company before deciding to purchase its stock to avoid buying risky stocks. This evaluation includes analysis of the company’s performance on the social media and financial news websites. However, such a huge amount of social media and financial news data cannot be completely assessed by investors. Therefore, an automated decision support system is necessary for investors, as this system will evaluate stock trends automatically using such large amounts of data. This automated system can be built using machine learning algorithms. Finding those algorithms that are more effective in predicting stock market trend using external data, like financial news and social media data or google trend etc. is very important. As accurate stock prediction based on external factors will increase investors’ profits, so machine learning researchers have taken a keen interest in this field. So I’m also building a platform that will be based on a machine learning model that will analyze the thousands of news and social media sentiments about a particular stock and predict weather the price of that stock is going to increase or decrease in near future.

Information Retrieval Machine Learning Natural Language Processing FinTech

Speech recogniton of Cerebral Palsy Patient

Cerebral Palsy (CP) is a group of disorders that affect a person's ability to move and maintain balance and posture. CP is the most common motor disability in childhood. Cerebral means having to do with the brain. Palsy means weakness or problems with using the muscles. It occurs during pregnancy or birth or within first 2 or 3 years of child (Staff, 2019). The project is about creating a device/app/environment which takes a patients impaired speech as an input and synthesize it into a comprehensible output speech.

Machine Learning Signal Processing Speech Recognition

Traveling Social and Aggregating Platform with a Recommendation System using Personality Profiling

A Travel Packages Aggregating platform and a Quora like community platform for travelers. Features like caching, Social Auth, Assistance through Emails, Selenium testing, and Rest API were integrated.

Web Development Data Scrapping

Breast cancer detection using deep learning

Breast Cancer is one of the leading causes of mortality among women aged 20-59 years. Its rate is increasing day by day, leading to a vast number of causalities per year. This rate can be reduced by early detection and treatment. Suppose the cancer is detected in its early stage. In that case, the severe consequences can be avoided, which takes an experienced physician to diagnose the breast cancer from a mammogram with approximately 79% accuracy. This accuracy can be increased up to 91% using machine learning techniques. There are several models for breast cancer detection which are based on machine learning. Some of those are more accurate, while others are less accurate while trained on the same dataset. We also need a comparative study to compare these models and evaluate their results regarding different bases (i.e., complexity, memory, application, and accuracy). To evaluate these models, they are trained on the same dataset CBIS DDSM: Curated Breast Imaging Subset of Digital Database for Screening Mammography by Lee et al. After evaluation of models best performing model is used to develop an application for medical experts (Android, React, and web).

Machine Learning Computer Vision

Detection of Nutrients Deficiency in Plants Using Machine Learning

Agriculture acts as the backbone of any country's economy. Low yield for crops poorly a ects the farmer, food chain, and economy of the country. One of the major factors for crops' growth rate reduction is nutrients de ciency or nutrients excess. Farmers get feared to approach agricultural experts for the detection of malnutrition and its treatment because it is an expensive process and time consuming as well. We are going to deliver a farmer-friendly mobile application that will take a close-up image of a plant's leaf as input, The image will be processed by a deep learning model, trained on some certain data set. The trained model serves its objective by providing information about nutrients. Then, AI (Arti cial Intelligence) Planning will play its role and will give a proper prescription for the treatment of malnutrition.

Machine Learning Computer Vision

Light Weight Voltage Transformation Unit

Weight and possible size reduction of Power transformer using Variable frequency drives to reduce the cross- sectional area of transformer’s core, eventually leading to smaller transformer with same power transformation capability.

Power Electronics

Designing of an Efficient Electrooculogram (EOG) Based Human-Computer Interface: Eye Controlled Cursor

We aim to design and implement a Human-Computer Interface (HCI) using the human eye to imitate the functions of the mouse (movement, Right/Left Clicks), fostering the automation and diminishing the physical effort by the human in operating the machines. Instead of using Camera and Image Processing, a non-invasive technique of Electrooculography (EOG) is aimed to cope with the constraints of facial expressions, environment and fix eye position in the technique of Image Processing. Using EOG cup electrodes, the cornea-retinal potential will be sensed, which after amplification and filtration will be fed to the computer to perform the corresponding action. The close to real-time movement of cursor will be decided by machine learning(ML) algorithm. Using a better version of clustering algorithms will be utilized to decide the movement of cursor corresponding to the cornea-retinal potential signal.

Embedded System Data Science Machine Learning HCI Biomedical Tech

Design and Control of Maximum Power Point Operation of Wind Turbine

Extent of this project basically incorporates planning of proficient framework for Wind Turbine through Doubly Fed Induction Generator(DFIG). Another finished result will be Power convertor based VFD which will be utilized to control the rotor frequency.

Renewable energy DFIG

Design of smart meter for monitoring and controlling bi-directional power flow

The main objective is to build up a single phase bi-directional which will be able to measure and control the power flow direction. The whole mechanism would be controlled by a microcontroller interfaced with calibrated circuitry.

Power Renewable Energy

Telemonitoring via Mobile Robot

The primary objective of our project is real-time monitoring of the environment via a mobile robot. The robot will be capable of moving from one place to another through a web interface for a safe and agile remote control. Additionally, a camera with 360-degree rotation will be mounted on the top of the robot for real-time monitoring of the environment. Raspberry pi / Arduino will be used to control the functionality of the robot. The developed product will make many things more accessible that will be efficiently used in monitoring space-limited and difficult environments.

Robotics Telemonitoring Embedded Systems

IOT enabled smart home with safety system

The objective of our project is to reduce the unnecessary use of electricity and the safety of our home (office) by controlling the device through the internet and automatically. This objective will be achieved by accomplishing the following tasks:
● Supply Voltage protection (200V-250V)
● Remote control of electric appliances
● Automatic tank overflow control
● Fire control
● Temperature measurement
● Display
● Webpage development or (Android App)

Internet of Things (IoT)

A Hybrid Renewable Energy Based Fast Charging Infrastructure for Electric Vehicles (EV)

The aim of this project is mainly to design A Hybrid Renewable Energy-based fast Charging Infrastructure for Electric Vehicle. As this technology reduces dependence on fossil fuels and decreases greenhouse gas emissions, electric vehicles (EVs) are becoming more common. Solar energy is a viable alternative source of renewable energy.

Power Electronics Renewable Energy Hybrid Renewable Energy

Sign To Speech

The project was inspired by the idea of converting signs to spoken language with the help of hand rotation and finger movements. The requirements for this project include a wearable glove that is responsive, portable, feasible, marketable, affordable, and durable. According to the bend angle value and the hand movement, through machine learning algorithm the Arduino microcontroller understands that which set of value represents which symbol, and then it will transfer the appropriate outcome value to the android app, which translates the text message to a voice message, via Bluetooth device, which produces an audible human hearing voice.

Automation Technology Sensors Arduino Embedded system

A highly Efficient on board charger with V2G capability for EV's

This is a Power Electronic based project in which converters and controller will be used. First Converter will be AC to DC mainly a rectifier and the second will be a Dual Active Bridge (DAB) DC to DC converter. Because the aim is to achieve efficiency so the soft-switching (it helps to reduce the switching loss because switching devices turn on and off at zero or almost zero voltage/current. Ploss=VI) techniques will be used to attain efficiency. Then the efficiency of the converter with and without soft switching will be compared to verify the effectiveness of soft switching techniques in terms of efficiency. And power flow is bidirectional so the bidirectional converters will be used to provide a bidirectional flow of power.

Power Electronics Transportation

Design of Power Converters for Harmonic Mitigation in Smart Micro Grid.

As a good solution to the energy crisis and environmental pollution, solar energy has been drawing more and more attention. We can provide solar energy to a smart microgrid through converters. The aim of this project is to design power converters that will convert the DC signal of solar PV into an output AC signal. Then the AC signal can be supplied to load and Grids. While providing AC energy to a smart microgrid through converters, harmonics create. These harmonics flow in the system and affect the source. So, an Active power filter will be used to reduce or eliminate these harmonics to increase solar PV lifetime.

Power Electronics

Soil Analysis using Machine Learning

We are going to provide a technical solution for soil Analysis at farms. Our end-product is an android mobile application. Farmers simply need to use our app and upload a picture of soil from their field. The image and their location are processed by the Machine Learning model which is trained on a certain dataset.

Machine Learning

Early Forest Fire detection using Internet of Things (IoT)

Our final year project is to design a forest fire detection system based on Internet of Things. The aim of this project is basically to control and prevent the forest fire in early stages so that most of loss of wildlife and log in fire would reduce. Their main role is to gather and cluster data from the node sensors, manage alarms and commands, and build the core network and to inform the control room about the fire in early stage, so that it could be prevented in time.

Internet of Things (IoT) Machine Learning

Fruits Online

This project shows the whole business plan of the online fruits store. The whole component of the business plan has been written. First the introduction of the idea is given. Then the characteristics of the fruit industry is written. In characteristics of fruits industry, the types of fruit market are given. Forces of the industry are studied. Proper environmental scan has been done in environmental scan the whole opportunities and threats have been calculated. On the basis of this the SWOT analysis has been done. Fruits Online is an online fruit store which deals with the most favorite and consuming fruits in Pakistan like Mangoes and Oranges. It is a partnership business between two persons. Initially we start as a small business and further have plans to expand it in future. Fruit Online deliver export quality fruit directly from their origin to customer doorsteps. We deliver via third party courier service and market our products on social media platforms. Behind the business is to grab the opportunity in the market which has a great potential. Everything is going to be digitalize and every trade is on E-commerce but fresh fruits are still on conventional patterns. To get a good quality fruit a person has to go market physically. It’s started in the Sadiqabad from the mango fruit and then extended to the oranges as well. Placement of this business is online. We don’t have any shop or store. We accept orders online only. Normally the fruit market covers five to six middle men and then its available to the customer. Due to large number of middle men the quality of fruits not remains enough good when it comes to the end user. The potential in the market is to minimize the middle men. The price will be automatically low when stake holders’ numbers will decrease. Conclusion and future ambitions are discussed to put the overall discussion into prescriptive.

Finance Business Management

The journey of organization plantation in the perspective CEO transformation leadership

HRM

Business plan on Board Games

Entrepreneurship is the process of evaluating an opportunity and then developing a product that will best serve that opportunity. The business plan has been created with this in mind. In this project whole business plan of Board Game (LUDO) is written that how a Ludo board gets delivered to retailers and how we can add features in existing product in order to gain competitive advantage. Personal selling, sales promotion and advertising technique will be used to target the customer. Previously, 180 products have already been sold making a profit of 9600. In the improved product, different branding techniques will be used such as displaying landscape photographs of famous places of related target market. A sample design is attached in the document.

Sales and Marketing

Compensation System Development and Use of Compensation Software

Compensation system is all about the financial and non-financial values given to the employees for the exchange of their services. The background of the project is to develop compensation system for hypothetical hotel which will be managed by information system. The hotel will consist of different position front office agents, housekeeping, food and beverage services, kitchen, engineering and maintenance, accounts, security, human resource, sales and marketing, purchase and information technology and in each position there is sub position. Through that system employees can easily access attendance, salary structure, and other additional benefits. For compensation system design we will use job evaluation in which we will determine the relative worth of the job based on compensable factors. In job evaluation we will be using ranking method, paid comparison method and point method. The compensation system will be designed through the compensation mix which includes base pay, performance pay and indirect pay. In base pay it includes job evaluation, market pricing and pay for knowledge. In performance pay it includes individual pay, group pay and organization pay. In indirect pay it includes benefits, pension plan and insurance etc. Nowadays there is a technological advancement due to which organization is shifting to computerized system in which all information is being managed through information system. For that purpose we will also be adding the component of information system for which we will be using Odoo. It is an ERP system in which we will develop a website which can be easily accessible from anywhere which contains all type of record like employee data, salary, benefits, attendance and working hours etc. When the overall compensation system is designed then in last we will add all these data in that website.

Enterprise Resource Planning (ERP)

Sexual objectification of women in advertisements increases acceptance of sexual aggression

Purpose of this study is to empirically substantiate if objectification of women in advertisements is increasing acceptance of violent sexual behavior among men in context of Pakistan. This study endeavors to deal with this concern. Due to the nature of the topic, and limited time frame, advertisements featuring women were selected and shown to selected sample of male adults. Detailed literature review and conclusions from previous studies, in relation to sexual objectification of women in advertisements, were discussed and pondered over and it was found that sexual objectification in advertisements contributes to a shift in attitudes and leads to acceptance of sexual aggression. A total of 150 female adults from different backgrounds participated in survey based quantitative study which used Likert scale. Furthermore, research recommends that admen should not manipulate the image of women to attain commercialized interests. Instead, admen and marketeers should exercise their positive role to make better ethical advertisements without objectifying females.

Marketing Advertisement

When and How Inclusive leadership is cultivated at work?

The purpose of this study was to examine the HPWS-inclusive leadership affecting employee’s behavior in hotel and hospital sector of Pakistan. Impact of HPWS-Inclusive leadership on innovative work behavior and in role performance with mediating role of Inclusive leadership was studied in detail. For this reason, a conceptual framework was developed and hypotheses were proposed to investigate the relationship among variables. Data collected from 180 respondents in different cities of Pakistan like Lahore, Islamabad, and Mianwali etc.

Human Resource Inclusive leadership High Performance Work System Inrole Performance Innovative Work Behavior

Financial Sustainability Model of Namal

This project is about sustainability model of Namal institute. Namal has a major issue of getting funds from donors. The purpose for this project is to make a forecasting model for the next 10 years. This will include all the ways Namal can be sustainable and generate funds of its own. This model will tell about all resources with which income can be generated. All of the universities especially in Pakistan doesn’t depend on one income resource which is Student fee. There are multiple income options a university can use to generate funds.
* Research
* Golden Scholarships
* Student fee
* Executive training programs
* Grant from government
* Artificial intelligence and agri tech department
I have collected all the data about Namal and my introduction part is complete. I am currently doing forecasting with the available data to measure 10 years data about Namal sustainability forecasting.

Finance

The impact of green vehicle advertising on the vehicle purchase intention: Moderating role of economic value by using planned behavioral theory

The purpose of this study was to investigate the impact of green vehicles advertising on the buying intention of cars consumers by using the planned behavioral theory. We used the economic value as a moderator variable which helped to know the economic prospect in advertising. The data of 200 plus was collected from car users mostly hybrid car users from all over the Pakistan and analyzes to present. This study will help the marketer to make economical products and advertiser to advertising agencies for designing the ad of green products.

Green Vehicles Advertising Economic Value PBT

Analysis report on selection criteria of companies

The selection criteria of any company is very important for it. Itis all the requirements for an employee to be eligible for the job. Different companies have different methods of recruitment for employee’s selection. For example, some companies select on the basis of the academics of candidates, some companies select on the basis of experiences and others companies like Tesla has selection criteria on the basis of skills. So, this has an important role for an employee’s job.

HRM