Fatemeh Askari Email and Phone Number

I am working on few-shot image classification. We have presented a model based on the prototypical network as its base. We have tried to improve the model by using multiple embedding spaces , weighting the feature vectors, and utilizing a self-attention mechanism.

Course: Artificial IntelligenceInstructor: Dr. Mohammad Reza MohammadiSemester: Fall 2023

Course: Graph theoryInstructor: Dr. Farzaneh Ghayour BaghbaniSemester: Fall 2023

I have worked on two projects:1)Tuberculosis Disease Diagnosis from X-ray Images:The dataset consisted of lung X-ray images categorized into two classes: whether the patient has tuberculosis (TB) or not. The main challenge we faced was the imbalance in the dataset. To address this issue, we employed techniques such as data augmentation and SMOTE(Synthetic Minority Over-sampling Technique). We fine-tuned the dataset using pre-trained models like ResNet, VGG, and DenseNet.2)The second project aimed to perform image similarity analysis using the image similarity metric. We had a dataset of flower images with 102 classes, and the goal was to train a model that could detect similarities and differences between the images. In this way, given two images, the model should be able to determine whether they belong to the same class or not. For this purpose, we allocated 82 classes to the training data, 20 classes to the validation data, and 20 classes to the test data. We created positive and negative pairs for the test data, where if a pair belonged to the same class, it was labeled as positive, and if not, it was labeled as negative.Next, we used a pre-trained ResNet model to extract feature vectors. We employed a loss function combining center loss and cross-entropy to encourage the feature vectors belonging to the same class to be close to each other in terms of Euclidean distance and to push away feature vectors from different classes. Finally, we evaluated the model on the test data using the ROC-AUC metric as our evaluation criterion.

Course: Algorithms Design & AnalysisInstructor: Dr. Marzie MalekimajdSemester: Spring 2023

Course: Data StructuresInstructor: Hossein RahmaniSemester: Fall 2022

Course: Advance Programming Instructor: Dr. Marzie MalekimajdSemester: Spring 2022

Course: Digital LogicInstructor: Dr. Hajar Falahati

Course: Discrete mathematicsInstructor: Dr.Vesal HakamiSemester: Fall 2021