About me
Hello! My name is Ghazal Bakhshande, And my nickname is Ghazal. I'm an undergraduate student pursuing my bachelor's degree in Computer Engineering at IUST with an outstanding academic record (GPA: 18.74/20).
My journey led me to the world of AI, where I've been exploring Deep Learning models, specializing in Computer Vision and Vision-Language Multimodalities.
I work as a computer vision research assistant under the supervision of Dr. Sauleh Eetemadi, an experienced researcher formerly at Microsoft. Currently, I'm engaged in the "SEMEVAL 2024 TASK 4: Multilingual Detection of Persuasion Techniques in Memes" project, focusing on multilabel classification and multimodal data analysis.
Experience
Academic Experience
Research Assistant at Computer Vision Lab, IUST
Image Segmentation
- Developed a Siamese U‑Net model for crack detection
- Optimized training with weighted binary cross‑entropy and dice loss
- Collaborated with real‑world tiles and patterns datasets
- Applied edge detection, polygon extraction, ORB keypoint‑based matching, histogram matching and perspective transformation
- Evaluated using dice coefficient, IoU, precision, and recall metrics
Research Assistant at NLP Lab, IUST
Vision-Language Multimodalities
- Analyzing State-of-the-Art Implementations in VQA Tasks
- Contributing to the ”SEMEVAL 2024 TASK 4: Multilingual Detection of Persuasion Techniques in Memes” project
- Engaging in Multilabel Classification and Multimodal Data Analysis
- Applying transformers and attention mechanisms
- Experimenting with Double Visual Textual Transformer, Visual‑Bert, and MultiModal BiTransformers
Teaching Experience
Computational Intelligence TA
- (Instructor: Prof. Nasser Mozayani)
- Sep 2023 - Present
Algorithms Design and Analysis TA
- Instructor: Prof. Marzieh Malekimajd
- Feb 2023 - Jul 2023
Software Engineering Mentor
- (Instructor: Prof. Behrouz Minaei‑Bidgoli, Prof. Mehrdad Ashtiani)
- Feb 2023 - Present
Database Design TA
- Instructor: Prof. Hossein Rahmani
- Feb 2021 - Jul 2021
Operating Systems TA
- (Instructor: Prof. Reza Entezari-Maleki)
- Sep 2022 - Jan 2023
Logical Circuits TA
- Instructor: Prof. Hajar Falahati
- Sep 2021 - Jan 2022
Fundamentals of Computer and Programming TA
- Instructor: Prof. Reza Entezari‑Maleki, Prof. Tayebe Rafiei
- Feb 2020 - Jan 2021
Computer Workshop TA
- Instructor: Prof. Marzieh Malekimajd
- Sep 2022 - Jan 2023
Advanced Programming TA
- Instructor: Prof. Tayebe Rafiei
- Feb 2021 - Jul 2021
Education
B.Sc. in Computer Science and Engineering
- Iran University of Science and Technology, Tehran, Iran
- Ranked 4th among Iran Universities based on QS Ranking (World Rank: 451)
- GPA: 18.74/20 (3.86/4)
- Last two years: 19.42/20 (4/4)
Diploma in Mathematics and Physics Discipline
- Motahare HighSchool, Tehran, Iran
- GPA: 4/4
Skills
Programming Languages
Python
C#
C++
HTML/CSS
VHDL
Matlab
FrameWorks & Libraries
Tensorflow
Keras
Pytorch
Django
Django-Rest
ASP.Net
Others
Git
Azure DevOps
SQLite
PostgreSQL
Xilinx ISE
ANTLR
PERSIAN
NativeENGLISH
TOEFL iBT will be taken soonSelected Projects
You can click on "View GitHub" to see more detailed description and implementation.
To see my other projects which are not mentioned here, visit my GitHub.
Tile Crack Detection with Siamese U-net
- Detecting tile cracks in images using processing techniques and deep learning.
- Implementing Siamese U-Net for the Segmentation Task.
Predicting COVID-19 From Chest X-Ray Images
- Predicting COVID-19 from chest X-Ray images using deep transfer learning.
- Utilizing Data Augmentation with a Pre-trained SqueezeNet Model.
CamScanner
- Implementing Contour detection, Perspective transform, and Morphology Operations for document processing.
- Applied various image processing techniques to emulate CamScanner filters.
Multi‑Feature Image Classification
- Implemented LBP texture feature extraction, shape characterization with compactness, eccentricity, and solidity metrics.
- Employed SVM for image classification on the Ships and Airplanes dataset.
Malware Detection
- Trained multiple deep models to classify malware vs. benign samples in a dataset.
- Developed Gradient Boosted Decision Trees and optimized thresholds using precision-recall curves.
Time Series Anomaly Prediction
- Developed three anomaly prediction models (RNN, LSTM, GRU) and innovated with self-supervised learning.
- Balanced data through sampling and improved accuracy using Min-Max scaling and L2 normalization.
Car Company Classification
- Car producer classification using deep transfer learning.
- Utilizing a pre-trained MobileNetV2 model and fine-tuning it.
- Employed Grad-CAM for model interpretability.
IOU-Based Face Detection
- Utilized image labeling tools and sliding window operations for generating proposals.
- Implemented Intersection over Union (IOU) for proposal classification
Radial Basis Function Approximation
- Implemented RBF Network with K-means, GMM, and Random clustering for function approximation.
- Trained an MLP to approximate the function and compared results with RBF models.
Berkeley CS188 Projects
- Berkeley CS188 Projects, Covering topics including Search Problems, Informed Search, Adversarial Search, MDP, and Reinforcement Learning (RL).
TSP Approximation with Kohonen Network
- Designed a Kohonen(SOFM) network to solve the NP-Hard TSP problem roughly close to the real answer.
Hopfield Neural Network
- Implemented a Hopfield network for pattern recognition, noise removal, and data recovery.
Genetic Algorithms
- Developed a genetic algorithm from scratch for the knapsack problem, with crossover and fitness functions.
Persian Optical Character Recognition
- Executed image pre-processing techniques to optimize Persian text extraction from images.
Mobile Data Tracker App
- Developed an Android app using Kotlin for monitoring and tracking mobile data usage.
- Allows users to set data usage limits and receive notifications when limits are reached
DoS Attack Detection System
- Simulated DoS attacks and developed a system to detect and mitigate them.
Student Hub Website
- Built a .NET Blazor website for students to rate professors, access course resources, and inform course decisions based on peer feedback.
Contact
Feel free to contact me at: bakhshande.ghazal@gmail.com
You can also reach me via these social media: