Milad Mirjalili
Department of Electrical Engineering, KNT University of Technology, Tehran, Iran
RESEARCH INTERESTS
- Machine Vision (Motion Analysis, Object Recognition, Action recognition, Augmented Reality)
- Machine Learning (Supervised Learning, Unsupervised Learning, Reinforcement Learning)
- Microcontroller Programming
- FPGA Programming
- Circuit Design
- Deep Learning (Convolutional Neural Networks, Generative AI, Generative Adversarial Networks)
- Image and Video Processing
Education
-
2018-2021
M.Sc. in Electrical Engineering
KNT University of Technology
GPA: 17.21/20 (3.77/4)
-
2012-2017
B.Sc. in Electrical Engineering
University of Shiraz
GPA: 15.27/20 (3.06/4)
RESEARCH EXPERIENCE
-
Developing a variable-size adaptive block matching motion estimation algorithm
▪ Using Python, I Created a method for block matching motion estimation (ME), with a focus on variablesize blocks. First, I designed a fast and efficient way to automatically generate blocks with different shapes and sizes according to the identified location and direction of movement. Next, by utilizing adaptive search tools, (e.g. the search window size selection, global motion compensation, changing reference frame), I attempted to decrease the computation load while maintaining the quality of ME.
- More details
- Supervisor: Dr. Amir Mousavinia (moosavie@kntu.ac.ir)
-
Implementing RGB to gray scale converter and Gaussian blur filter using VHDL
▪ I implemented RGB to grayscale converter and Gaussian blur using VHDL. First, I calculated a weighted sum of image pixels to create a grayscale image. Second, I designed the filter by implementing the process of convolution by shifting a Gaussian kernel on an image.
- More details
- Supervisor: Dr.Hossein Hosseini-Nejad (hosseini_nezhad@kntu.ac.ir)
-
Designing digital to analog converter (DAC) and Analog to digital converter (ADC) using Cadence
▪ Using Cadence in two different projects, I implemented 10-bit current steering DAC and 5-bit Flash ADC in CMOS, 0.18μm technology. I selected the width and length of transistors based on the required design criteria, such as spurious-free dynamic range (SFDR).
- More details
- Supervisor: Dr.Hossein Shamsi (shamsi@eetd.kntu.ac.ir)
-
Designing a fully differential two-stage op-amp and drawing the layout using Cadence
▪ Using Cadence, I designed a fully differential two-stage op-amp with PMOS transistors as inputs in CMOS, 0.18μm technology, and drew the layout. The gain, slew rate, and phase margin are above 50 dB, 94 V/μSec, and 60°, respectively.
- More details
- Supervisor: Dr. Hossein Shamsi (shamsi@eetd.kntu.ac.ir)
-
Designing a class E power amplifier using Cadence
▪ Using Cadence, I implemented a class E power amplifier in CMOS, 0.18μm technology. I chose the components based on design criteria. For instance, the output power is 100 mW, and the frequency is 100 MHz.
- More details
- Supervisor: Dr. Mehdi Ehsanian (ehsanian@kntu.ac.ir)
-
Designing and creating a portable solar battery charger using AVR
▪ Using AVR, I designed and created a portable battery charger containing three solar cells arranged in parallel. I used the LM2576 regulator to bring the panels’ voltage and current to the desired level. I also used ATMEGA8 to display current and voltages on an LCD.
- More details
- Dr. Mehdi Miri (miri@shirazu.ac.ir)
-
Designing and creating a signal jammer
▪ In this project, I aimed to create a circuit to block undesired signals. The selected range of frequency was between 80 to 110 MHz. I used BC337 transistor to amplify the signal and a variable inductor and capacitor to adjust frequency.
- More details
- Dr. Mehdi Miri (miri@shirazu.ac.ir)
-
Simulating the performance of a PIN diode using Silvaco
▪ I aimed to simulate the function of a PIN diode at different voltage and doped levels using Silvaco. I analyzed the effect of different variations in doping and voltages in the operation of this device by inspecting various plots, such as hole concentration and electric field.
- More details
- Dr. Mehdi Miri (miri@shirazu.ac.ir)
OTHER PROJECTS
-
Software for Optical Character Recognition (OCR)
▪ I implemented a commercial program to extract text information from catalogs. I used OpenCV to extract regions of interest and preprocess images. For example, I used a series of morphological transformations, such as the opening to detect horizontal and vertical lines. Finally, I used tesseract for OCR.
-
Software for managing patients’ medical insurance information
▪ Using Python, I developed a commercial software for registering patients’ info. The user interface was built using the PYQT5 library, and the data storing and reporting were done using the sqlite3 module and Pandas library
-
Software for extracting speaking parts from a movie based on the subtitle file
▪ I developed a commercial software for extracting talking parts from a movie based on subtitle timing and generating new subtitle files for that new video. Various encoding settings can be modified, such as using HEVC or AVC.
PRESENTATIONS
-
2020
- Adaptive Block Matching Motion Estimation, Mirjalili M
- Presentation video
- Presentation slides
- Recognized the best seminar in 5th Student Seminar on Electrical and Computer Engineering Innovation 2020
-
2021
- In this video, I discuss the fundamentals of bitcoin mining.
-
2021
- In my YouTube channels, I tried to show different practical usage of computer vision in real life, such as colorizing black and white images, object detection using haar features, and creating a virtual pen using OpenCV.
PRESENTATIONS
-
2023
- “Efficient Block Matching Motion Estimation Using Variable-Size Blocks and Predictive Tools” (Submitted), Mirjalili M, Mousavinia A
-
- “Deep Learning-Based Approach for Optimal Block Size Determination in Block Matching Motion Estimation” (In Preparation) , Mirjalili M, Mousavinia A
INTERNSHIPS
-
2016
-
Intern at Khane Sakhtafzar, Shiraz, Iran
▪ I learned to read computer circuits schematics, how they’re operating, and testing different electrical components. Additionally, I got familiar with the safety protocols during working in an industrial environment.
-
Intern at Khane Sakhtafzar, Shiraz, Iran
-
2014
-
Intern at FJT Institute in collaboration with Shiraz University, Shiraz, Iran
▪ I learned to design solar panels for industrial and home use and their safe installation.
-
Intern at FJT Institute in collaboration with Shiraz University, Shiraz, Iran
COMPUTER SKILLS
- Programming: Python, MATLAB, C, VHDL, Java
- Libraries: OpenCV, Pandas, TensorFlow, scikit-learn, NumPy, Matplotlib, Seaborn, SciPy
- Software: Cadence, Xilinx ISE, Android Studio
LANGUAGE
- English: TOEFL iBT score: 100 (Reading: 28 Listening: 27 Speaking: 21 Writing: 24)
- Persian: Native
CERTIFICATIONS
-
2022
- Advanced Computer Vision with TensorFlow, Coursera
-
2022
- Convolutional Neural Networks , Coursera
-
2022
- Neural Networks and Deep Learning , Coursera
-
2016
- Electrical components testing and repairing , Khane Sakhtafzar
-
2014
- Designing and installation of solar panels , FJT Institute
REFERENCES
- ▪ Master’s Thesis Supervisor | Dr. Amir Mousavinia moosavie@kntu.ac.ir +98 -21-84062228
- ▪ Bachelor’s Thesis Supervisor | Dr. Mehdi Miri miri@shirazu.ac.ir +98 -71-36133190
- ▪ Course Project Supervisor | Dr. Hamid Abrishami Moghaddam moghaddam@kntu.ac.ir +98 -21-84062229