Projects

Forecasting of sales for the Walmart company (From EDA to forecasting + documentation)

We have undertaken an extensive benchmark evaluation of LLMs and conversational bots. Our evaluation involved the collection of 1002 questions encompassing 27 categories, which we refer to as the “Wordsmiths dataset.” These categories include reasoning, logic, facts, coding, bias, language, humor, and more. Each question in the dataset is accompanied by an accurate and verified answer. We meticulously assessed four leading chatbots: ChatGPT, GPT-4, Bard, and Claude, using this dataset.

Persian character-level poetry generation using LSTM (PyTorch)

Sentiment analysis for Persian comments using FastText.

What happens if we encounter a suitable font for our design work but do not know its name? Visual Font Recognition (VFR) systems are used to identify the font typeface in an image. These systems can assist graphic designers in identifying fonts used in images. A VFR system also aids in improving the speed and accuracy of Optical Character Recognition (OCR) systems. In this paper, we introduce the first publicly available datasets in the field of Persian font recognition and employ Convolutional Neural Networks (CNN) to address this problem. The results show that the proposed pipeline obtained 78.0% top-1 accuracy on our new datasets, 89.1% on the IDPL-PFOD dataset, and 94.5% on the KAFD dataset. Furthermore, the average time spent in the entire pipeline for one sample of our proposed datasets is 0.54 and 0.017 seconds for CPU and GPU, respectively. We conclude that CNN methods can be used to recognize Persian fonts without the need for additional pre-processing steps such as feature extraction, binarization, normalization, etc.

Apply different colors to your hair! The next update would be about applying lipstick. The following technologies are used for this project:

• CNN architecture: U-Net
• Dataset: CelebAMask-HQ
• Framework: TensorFlow

The process consists following steps:

• step1: upload your image
• step2: apply image segmetation on your image/video to generate mask image
• step3: apply color transformation on your original image/video based on mask image
• step4: enjoy!

• Iris problem (Dense, classification)
• Fashion-MNIST problem (Dense, classification)
• CIFAR-10 problem (CNN, classification)
• Fashion-MNIST problem (CNN, classification)
• Handwritten Digits Recognition problem (CNN, classification)
• Learning Rate Finder
• Write Coustom CallBack
• Using pretrainded model
• Flowers problem using tf.data (CNN, classification)
• Cat vs Dog problem using tf.data (CNN, classification)

Detect Parkinson’s disease just from a wave sketch image. I provied a method by using Few-Shot learning to detect Parkinson based on the proposed dataset in paper.

Search between the objects of an image using YOLO and CLIP. You give a prompt about an object in an image, then the model will cut the region of the detected object based on your prompt.

A recommendation system for images (specifically painted portraits), for Web Mining and Social Networking Course.

• Architecture: MobileNet
• Framework: TensorFlow

Implementation of the Canny Edge Detector algorithm using MATLAB.

Do you want to run CenterNet pre-trained models in Colab without bugs and make predictions? This project is a fixed version of the CenterNet specifically for Google Colab.

A deep learning model for predicting the anxiety level of a person based on the Taylor Manifest Anxiety Scale. This test was published by Taylor, J.A in paper entitled“A personality scale of manifest anxiety (1953).”

Number Guessing Game by using ATMega32, Keypad, and LCD (C, Proteus, Code Vision AVR)

• Semester: Fall 2021 - Islamic Azad University, Mashhad branch
• Lecturer:‌ Dr. Mehrdad Jalali
• Website: https://mehrdad-dev.github.io/webmining-course-fall2021/
• GitHub Repo: https://github.com/mehrdad-dev/webmining-course-fall2021
• Tools: Python, RapidMiner, Gephi

Home Work list:

• Homework 1: applications, advantages and tools of the web mining
• Homework 2: Graph Partitioning Clustering - Terms (maximal clique & Single Link Clustering Method)
• Homework 3: RapidMiner - Classification and Clustering on a kaggle dataset about phishing websites
• Homework 4: KNN on webpages - N=3
• Homework 5: Graph Partitioning Clustering - Pages (maximal clique & Single Link Clustering Method)
• Homework 6: User Classification - KNN
• Homework 7: Review an article about the recommender systems
• Homework 8: Predict item score for a user in recommender system
• Homework 9: Collaborative Filtering Recommender System For Movie (MovieLens Dataset, Python)
• Homework 10: Item-based collaborative filtering
• Homework 11: Analysis of a graph network using Gephi

My Presentation: How to Find Maximal Cliques in a Graph using Python & ​KNN Classification example using scikit-learn library

• Semester: Fall 2021 - Azad University, Mashhad branch
• Lecturer:‌ Dr. Vahid Torkzadeh
• GitHub Repo: https://github.com/mehrdad-dev/intro-computer-vision-course
• Tools: Matlab (R2019b)

Published notes and translated the content while completing this course.

Jami is a simple cross-platform GUI for playing some musical instruments. By using Jami, you can play violin and piano easily on your laptop! Jami now is in the fyne apps list!