• 🔭 I’m currently working on Medical Image Processing

• 👯 I am cooperating with ISMILE company in Netherland Medical image processing

• 👨‍💻 All of my projects are available at https://github.com/RojaGhasemi

• 📫 How to reach me [email protected]

• 📄 Know about my experiences in my Linkedin https://www.linkedin.com/in/roja-ghasemi-94020974?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=ios_app

Here are some of my notable projects:

• Description: Brain Tumor Classification and Detection Using Combination of Deep learning and Machine Learning from CT Scan Images.
• Technologies: Python, Jupyter, TensorFlow, OpenCV , Keras.
• Status: Under submission in Journal.

• Description: Distinguish Between Brain Tumor and Hemorrhage from Ct Scan images using Machine Learning and Deep Learning models (use Transfer learning, pe-trained modesl and fine-tunnig)
• Technologies: Python, Jupyter, PyTorch,TensorFlow, OpenCV , Keras, Pythorch.
• Status: Under submission in Journal.

• Description: Distinguish Between Brain Tumor and Stroke from Ct Scan images using Deep Learning models (Use Transfer learning, pe-trained modesl)
• Technologies: Python, Spyder, PyTorch,TensorFlow, OpenCV , Keras, Pythorch.
• Status: Under submission in Journal.

• Description: This research proposes a CNN-based architecture for classifying fashion-MNIST photos using convolutional neural networks. The fashion-MNIST dataset includes 28x28 grayscale pictures of 70,000 fashion items from 10 classifications. The model's accuracy exceeded 93% in experimental results.
• Technologies: Python, Jupyter, PyTorch,TensorFlow, OpenCV , Keras, Pythorch.
• Status: Under submission in Conference.

• Description: The proposed method involves tumor segmentation, feature extraction, and classification using Zernike moments combination and discrete wavelet transform. The ACO algorithm selects useful features, and SVM cluster classifies them for tumor diagnosis. Fuzzy logic is used for tumor area segmentation. Experiments show efficiency and accuracy, with SVM having 88% accuracy.
• Technologies: Matlab.

• Description: This article presents an automatic method for classifying mammography images into normal, benign, and malignant categories using Gabor filters and LSD analysis. Tested on DDSM database data, it outperforms K-NN classification algorithms, suitable for clinics and hospitals.
• Technologies: Matlab.

• Description: The study employs a combination of DWT, local binary pattern, Zernike moment combination, and gray level co-occurrence matrix for feature extraction in medical imaging. These techniques provide faster calculations, detailed analysis, and high recognition accuracy in K-NN classification, with results on MIAS and DDSM databases.
• Technologies: C++, OpenCV.

• Description: The research introduces a new algorithm for classifying skin images into malignant and benign types, achieving an accuracy of 89%, making it a safe and effective method for skin cancer detection.
• Technologies: Matlab.

• Description:This paper presents a novel facial expression recognition method using locally weighted Pseudo Zernike Moments (LWPZM) and fuzzy classification. Extensive experimental investigation using the Radboud Faces database shows significant improvement in the method compared to other methods.
• Technologies: Matlab.
• Status: Publishe in IEEE: https://ieeexplore.ieee.org/document/6675658/authors#authors

• Description: This article proposes a new system using Fuzzy logic for facial expression recognition. The system uses a novel method for facial region extraction from static images, determining effective areas from integral projection curves. The system is robust, accurate, and superior to other approaches, achieving superior results on the JAFFE database.
• Technologies: Matlab.
• Status: Publishe in IEEE: https://ieeexplore.ieee.org/abstract/document/6802544/authors#authors

• Description: Diagnosis of brain tumor based on moments of local binary patterns and discrete wavelet transform from MRI images
• Technologies: Matlab.

• Description: In this research, an automated method for finding the exact range of the abnormal brain region in an MRI image has been investigated. This algorithm is based on fuzzy fragmentation that is used to detect a tumor in a brain medical device. The fuzzy fragmentation algorithm has the ability to quickly diagnose areas with heterogeneous and asymmetric tissue in a medical image. In order to extract the feature from the Gabor filter, it is a powerful method for extracting image texture. The PCA algorithm has been used to select useful features. In the end, SVM has been used to categorize the extracted features and tumor detection.
• Technologies: Matlab.