Thesis (Bachelor of Science in Computer Science)

Browse

End date: 2019

Search Results

Now showing 1 - 10 of 77
  • Thumbnail Image
    Item
    A quantum Z-transform
    (BRAC University, 2019-12) Sarker, Md. Sajibur Rahman; Majumdar, Mahbub Alam; Subramanian, Sathyawageeswar
    We investigate a quantum analog of the classical Z-transform with the aim of making it implementable on quantum computers, potentially offering a speedup over the classical method. Unlike the discrete Fourier transform, which is limited to frequency analysis, the Z-transform allows for versatile exploration of properties within the complex plane. Since the quantum Fourier transform underpins Shor’s factoring algorithm and serves as a subroutine in many other quantum algorithms, a quantum Z-transform promises broad applicability in quantum simulation, quantum machine learning, and quantum signal processing. This is especially relevant because Z-transforms generalize Fourier transforms in certain aspects. Given that quantum computers are particularly adept at performing unitary operations, we discretize the classical definition of the Z-transform and unitarize its matrix formulation to make it amenable for quantum computation. Our approach involves introducing a discrete Z-transform, mapping the input sequence to a discrete set of values to represent them as quantum states, and redefining the Z-transform as a finite summation to effectively handle the infinite summation of the classical definition. We then develop a matrix formulation for our redefined discrete Z-transform and extend our approach by unitarizing this matrix formulation through block-encoding, constructing unitary operators that meet the criteria for efficient quantum operations using standard quantum gates and subroutines. Our approach establishes the groundwork by fulfilling the mathematical foundations for the potential discovery of a quantum Z-transform and opens avenues for further exploration and implementation in quantum computing.
  • Thumbnail Image
    Item
    Detection of handwritten text using convolutional neural network
    (BRAC University, 2019-04) Jasim, Rabib Bin; Mahin, Rokeya Sultana; Uddin, Jia
    Machine replication of human functions, like reading, is an ancient dream. However, over the last five decades, machine reading has grown from a dream to reality. We have tried to make it more obvious through a hand writing recognition system. This research paper describes a text-line extraction based method. It offers a new solution to traditional handwriting recognition techniques using concepts of Deep learning and computer vision. An image can have hand writing, typed letters, different characters and other images. Our intention is to detect all the characters and display them. Some images can also have unnecessary lines or unclear letters. This system will clear the picture through pre-processing system and will be able to identify the letters or characters. It will help people to identify any unclear messages. It will also avoid unnecessary images and will focus on the text only. Sometimes we want to ignore unnecessary advertisement images from the newspapers. Our system will do a great work for this. It will clear all the images and unnecessary lines etc. and will only display the text what people want to read.
  • Thumbnail Image
    Item
    A new multi robot search algorithm using probabilistic finite state machine and Lennard Jones potential function
    (BRAC University, 2017) Khan, Md. Shadnan Azwad; Hasan, Mohammad S.; Ahmed, Tarem
    Swarm robotics is a decentralized approach to robotic systems. This paper exammes the problem of search and rescue using swarm robots. We present as solution a multi-robot search algorithm using probabilistic finite state machine and interaction inspired by Lennard-Jones potential function. The approach utilizes a finite state machine to separate the tasks performed and to change coordination rules according to the circumstances and social probabilities. The approach is tested in various scenarios to test flexibility, scalability and robustness. The performance results are promising and comparison with Robotic Darwinian Particle Swarm Optimization and Glowworm Swam Optimization for algorithmic complexity appear favourable.
  • Thumbnail Image
    Item
    Indoor positioning techniques using RSSI from wireless networks
    (BRAC University, 2019-08) Sohan, Asif Ahmed; Fairooz, Fabiha; Rahman, Adham Ibrahim; Ali, Mohammad; Chakrabarty, Amitabha
    The whole world is familiar with the Global Positioning System or GPS which can identify the exact position of any object with the help of satellites. Yet GPS signals are not available indoors. To overcome this, Indoor Positioning System(IPS) is used which enables us to locate objects inside an indoor environment. Our goal is to build an Indoor Positioning System by estimating the location using Received Signal Strength Indication (RSSI) through wireless networks. The proposed model will determine the position of wireless devices in a room. We took the RSSI values as coordinates and speci c reference points at every two meters making the room into a grid. The RSSI values on the reference point are measured. The position of the wireless devices will be estimated from the reference points using the trilateration method and the ITU indoor path loss model. With the aforementioned process, we calculated the position using the ITU indoor path loss model and trilateration. Using the ITU indoor path loss model our mean error was 1.01166m and while using trilateration it was 1.22m.
  • Thumbnail Image
    Item
    Comparative study of X-ray and CT scan images for the detection of COVID-19 using deep learning
    (BRAC University, 2015-08) Niloy, Ahashan Habib; Shiba, Shammi Akhter; Fahim, S.M. Farah Al; Faria, Faizun Nahar; Rahman, Md. Jamilur; Parvez, Mohammad Zavid
    Coronavirus 2019 (in short, COVID-19), originated in the Wuhan province of China in December 2019, has been declared a global pandemic by WHO in March 2020. Since its inception, it’s rapid spread among nations had initially collapsed the world economy and the increasing death-pool created a strong fear among people as the virus spread through human contact. Initially doctors struggled to diagnose the increasing number of patients as there was less availability of testing kits and failed to treat people efficiently which ultimately led to the collapse of the health sector of several countries. To help doctors primarily diagnose the virus, researchers around the world have come up with some radiology imaging techniques using the Convo lutional Neural Network (CNN). While some of them worked on x-ray images and some others on CT scan images, none worked on both the image types. Thus there’s no way to know which image works better for a particular model. This, therefore, insisted us to perform a comparison between x-ray and CT scan images. Thus we came up with a novel CNN model named CoroPy which works for both the image types and shows that in 2 classes (normal and covid), CT scan images show a better accuracy and it is 99.17% whereas it is 95.73% for x-ray images. However, in the case of 3 classes (normal, covid and viral pneumonia), x-ray images show a better accuracy and it is 92.45% whereas it is 68.81% for CT scan images.
  • Thumbnail Image
    Item
    Prediction of acute lymphoid leukemia using Privacy Preserving Neural Network
    (BRAC University, 2019-12) Khilji, Ishfaque Qamar; Saha, Kamonashish; Shonon, Jushan Amin; Israq, Ragib; Hossain, Muhammad Iqbal
    In today’s world, machine learning has become a big factor. It not only needs to be helpful, but also accurate and precise prediction is required. Machine learning is now becoming a widely used mechanism and applying it in certain sensitive fields like medical and financial data has only made things easier, but it also brought some difficulty in data privacy and data security which will protect the complete implementation of cloud based machine learning for these aspects due to the law and ethical needs. In this project, to give proper solution, we have come up with the idea using concepts of CryptoNets and Neural Networks, where we will be able to convert the learned neural network with the encrypted data to Cryptonets and the data will be totally encrypted and this will prevent the chances of unencrypted data being available to everyone. In this method, the owner will send the encrypted data to the cloud first and will hold a private key which can be used to decrypt the data later on. The cloud will have no idea about the data there since it will be in encrypted form and any attempts to get data from the cloud will only give the encrypted form. However, applying neural network to the cloud will enable us to store the data and make predictions in encrypted form and also give back the encrypted data to the user. In this way, the cloud will have no idea about the actual data and after the prediction is made, it will give back the predicted data in the encrypted form. We were able to achieve an encrypted prediction of about 78% close to the validation accuracy amount we achieved when training our Neural Network model.
  • Thumbnail Image
    Item
    A macroeconomic model for forecasting crude oil prices with Feedforward Neural Network Grid Search Experimentation
    (BRAC University, 2019-12) Aunjum, Md. Ragib; Naqi, Muhammad; Jamil, Sifat; Majumdar, Mahbubul Alam
    Crude oil is one of the most important determinant of the global and national economy and important decision making factors of industrial activities. For this reason, numerous mathematical and machine learning approaches have been conducted to predict the future trend of oil market. Yet, to predict the price of oil is one of the most challenging issues out there because the high volatile nature of oil market and the dependency of price on other factors. In many approaches on predicting oil price use machine learning algorithms, the only factors considered are the opening and closing prices. Thus, the implementations did not reflect the price pattern truly and also hampered the sudden ups and downs of price because the oil market does not only depend on the daily pricing behavior. By reviewing the historical data of oil market it can clearly be seen that the oil market is heavily affected by the geopolitical, technical and macroeconomic factors. For example, geopolitical factor such as war in middle east made the oil price soared high and broke the pattern of daily fluctuations by a large margin. And also, it can easily be seen that the everyday demand of oil along with the quantity supplied affects the oil market. So, these factors along with other macroeconomic and technical issues must be addressed to successfully determine the oil price trend. To justify our claim, we approach to predict the oil price using only the opening and closing market price by ARIMA, SVR and Linear regression model. Afterwards, the macroeconomic, technical, geopolitical factors were considered to predict oil price using Feed Forward Neural Network and compared the results with the ones we have found on the previous models.
  • Thumbnail Image
    Item
    A supervised learning approach by machine learning and deep learning algorithms to predict type II DM risk
    (BRAC University, 2019-09) Farabe, Abdullah Al; Sharika, Tarin Sultana; Raonak, Nahian; Ashraf, Ghalib; Chakrabarty, Amitabha
    The application of Arti cial intelligence (AI) has become a valuable part of medical research. These days diabetes is one of the top maladies on the planet. Nowadays it has become a common disease and alarming as people are living in polluted areas and eating unhygienic foods. People with diabetes are probably going to pass on at a more youthful age than individuals who don't have diabetes. We hope this study could be very helpful in medical science to predict the risk score of type II Diabetes Mellitus (DM). Our model consists of four machine learning algorithms which are- K-Nearest Neighbor, Random forest, Decision tree and Logistic Regression. These algorithms have been applied on a dataset containing 15000 type 2 diabetes patients along with eight features that describe the state of patients such as glucose, BMI, age, pregnancy, blood pressure (BP), Diabetes Pedigree Function, Skin thickness and insulin. Moreover, one deep learning algorithm called CNN has been used. All of the ve algorithms have been used on the dataset and the Random forest gives the best accuracy of almost 92.60 percent where other algorithms give less accuracy.
  • Thumbnail Image
    Item
    Personal information from Bangla speech signal using MFCC and GMM
    (BRAC University, 2019-08) Hridy, Maisha Munawara; Hasan, Md. Hasib; Emon, Mahfuz Al; Uddin, Jia
    Our system extracts personal information from bangla speech. Dataset that was used consists real-life voice inputs from di erent age and gender groups. A set of Bengali speech samples from YouTube were used as input dataset. This system is based on basic machine learning algorithms. Mel frequency cepstral coe cient was used to train and construct this system. While calculating gender and age detection part, we will be using GMM to calculate the nal scores on the samples having the MFCCs of the extracted speech samples. GMM model basically congregates some subsets among the whole set based on probability. Along with the gender determination process, age detection process will also be simulated using fundamental frequency of speech. Python is the programming language used to write the coding. Our system was successful in giving 88% accuracy for gender recognition and 75% accuracy for age detection.
  • Thumbnail Image
    Item
    Secured iota enabled micropayment transaction crypto-platform with discretionary mining capabilities and miner nomination based on first-price sealed bid auction theory
    (BRAC University, 2019-12) Amin, H.M Sadman; Rafi, Nakhla; Sufiyan, Zahin; Anjum, Syeda Afrida; Alam, Golam Rabiul
    The commercial utilization of cryptocurrency as a digital asset in being more and more sought-after on each successive year. Most of the well renowned cryptoplatforms now-a-days are devised based on the premise of the concept of blockchain. These cryptocurrencies work as an alternative medium of exchange using cryptography to secure the transactions on a distributed ledger. For validating a rationally substantial transaction, all of the blockchain-based cryptocurrencies requires a miner who will execute the proof-of-work to secure network consensus even in the presence of malicious nodes. After solving the cryptographic hash puzzles, the mining nodes are then compensated with a block reward and a mining/transaction fee for their services. When using regular blockchain-based digital cryptocurrency such as bitcoin, ripple, etherium etc. most of the time the respected crypto-platforms discourages low valued transactions from being executed. Because often, the transaction fee may exceed the value of the product or service that are being purchased. As a result, for this particular reason, micropayment systems using digital crypto-platforms remains largely under developed. To solve this complication our thesis model was emanated from the notion of IOTA, which is considered as a minerless cypto-platform where the requisition of miners are disregarded thus enabling users to relish the advantages of microtransactions, however with the inclusion of “Discretionary mining”. “Discretionary Mining” refers to the hypothesis of availability of mining capabilities at the discretion of the users. While using IOTA, the efficiency of the Tangle network largely depends on the computational power of nodes. Moreover, unconfirmed transaction node also known as “Tips” with low computational capability may not be able to validate it’s previous two transactions in time which will result in the degradation of the entire Tangle. Hence, our research of Discretionary Mining was derived from the postulation of distributive computing where a low powered smart device can outsource complex computations while validating transactions such as solving cryptographic puzzle (Hashcash) which they cannot execute on their own. Thus enabling users to reap the benefits of microtransactions without the network being completely minerless.