zahra shahlaie's Projects
Artificial Intelligence Fundamental Concepts
evaluate the capability of CNNs in detecting and classifying monochrome images based on different colors , the project evaluates the network's performance on noisy data to understand its robustness against noise and whether it can maintain its performance under such conditions.
This project aims to develop a deep learning model for detecting cracks in reinforced concrete structures under low light conditions and night simulation. Additionally, images were taken from further distances without zoom, and various textures were added to the images to improve the model's accuracy in crack detection.
This project involves training a facial emotion recognition model using transfer learning with pre-trained convolutional neural networks (CNNs). The code is designed to handle imbalanced datasets, perform data augmentation, and visualize training results. The models evaluated include VGG16, VGG19, ResNet50, and MobileNetV2, along with a custom CNN
Detecting facial emotions using YOLOv8 model and deep learning techniques
a clustering method using the heat and wave equations along with eigenvalues and eigenvectors. This approach is capable of clustering graphs into two partitions with a precision similar to spectral clustering but with faster execution time and lower computational complexity.
Human Image Segmentation with U-Net Architecture: This repository contains code and notebooks for a project focused on segmenting human images using the U-Net architecture. The goal is to accurately identify areas where humans are standing and generate corresponding segmentation masks.
This project aims to classify input samples of the Ronash dataset into 20 categories based on the title, vendor, and tags. The updated version of the multi-label text classification model, with applied improvements, achieved an accuracy of 97.86% and reduced the loss to 0.1043.
This repository contains my responses to the practice questions from the Artificial Intelligence course conducted by IPM. In this course, we were required to answer a set number of questions, and I have prepared my responses in the form of Jupyter notebooks.
This project involves creating a dataset for training a YOLOv8 object detection model by generating composite images of various input images overlaid on different background images.
This project demonstrates how to transform raw text data into word embeddings that provide a deeper semantic understanding of words and their relationships.
Config files for my GitHub profile.