No description available

Repository processes CT scanned images of human Lungs , which are in DICOM image format. Visualises the data in 3D and trains a 3D convolution network on the data after preprocessing.

In my LUNA16 dataset cancer detection project, I utilized PyTorch to develop and train convolutional neural networks (CNNs) for accurate identification of lung nodules in 3D CT scans. By leveraging advanced deep learning architectures and extensive data preprocessing techniques, the model achieved high precision and recall.

Lung Cancer Image Classification is a deep learning project that uses convolutional neural networks (CNNs) to analyze CT scans for early detection of lung cancer. Implemented in a Jupyter notebook, it features data preprocessing, model training, and visualization using the Lung and Colon Cancer Histopathological Image Dataset (LC25000).

基于matlab的PRED+CT数据集处理程序

This repository processes CT scan images of human lungs available as DICOM image format. It visualizes the data in 3D and trains a 3D convolutional network on the data after preprocessing.

This project develops a deep learning system to classify brain CT scans as normal or hemorrhagic using MobileNet V2. It processes a dataset of 71,491 DICOM images, achieving ~90% accuracy. The system includes a preprocessing pipeline for efficient data handling and is designed for real-time clinical use.

Most of the people (87%) are affected by Ischemic stroke. To distinguish this hemorrhage, we got to do a lab test at therapeutic. But that's exceptionally costly for our country Individuals. So, we choose to do something for them. After huge research, we developed a project which can detect this type of hemorrhage using deep learning neural network based algorithm. To do that we collected raw data (CT Scan Copy) from a number of hospital, we preprocessed it with a help of radiologist and finally trained our model to nail our goal, which is detecting hemorrhage stroke. Right now, we have achieved 90 up percentage accuracy in total.

A complete PyTorch implementation of a 3D U-Net for liver segmentation from CT scans. Features advanced preprocessing (CLAHE, Gaussian Blur), 3D data augmentation, and a robust training pipeline.

"This project predicts hospital stay duration using healthcare infrastructure data such as MRI units, CT scanners, and hospital beds. Regression models were trained on both raw and preprocessed data, and evaluated using R², MAE, and RMSE to assess the impact of preprocessing on prediction accuracy.

A deep learning project for slice-level classification of colon cancer from CT images, incorporating preprocessing, data augmentation, transfer learning, and Grad-CAM-based visual explainability within a doctor-in-the-loop framework.

Official code release for the Scientific Data publication “A spatially and temporally aligned contrast–non-contrast cardiac CT dataset of pigs.” Includes preprocessing, augmentation setup, nnUNetv2 training configuration, and cross-validation scripts for reproducible cardiac segmentation on CECT and NCCT data.

Radiomics and machine learning pipeline for predicting brain metastasis recurrence after radiotherapy. Includes CT/MRI preprocessing, mask generation, feature extraction, clinical data integration, and model training/evaluation. Developed as part of a Master’s thesis.

Lung cancer detection project using CNN and EfficientNetB3 to classify CT scan images. The model applies advanced feature extraction to improve accuracy and support early diagnosis. Includes data preprocessing, training pipeline, evaluation metrics, and visualization of predictions.

Worked with a team to implement a deep learning model to Identify kidney tumors in CT scans with an accuracy higher than the industry standard of 83%, preferably achieving around 75%-80%. I was responsible for data preprocessing.

Hybrid deep learning model for stroke prediction combining CNNs for spatial feature extraction from CT/MRI scans and LSTMs for temporal analysis of clinical time-series data. Achieves 94%+ accuracy through multimodal data fusion. Includes preprocessing pipelines, training scripts, and deployment-ready inference code.

A deep learning-based system for multi-class medical image analysis. This project uses CNN-based models to classify medical images (X-ray/MRI/CT) into multiple disease categories. It includes data preprocessing, model training, evaluation, and visualization of results to support faster and more accurate clinical diagnosis.

Developed a Deep Learning pipeline using the U-Net neural network architecture to automate the detection of fibrosis in cardiac CT scans. I managed the entire model lifecycle: from preprocessing raw DICOM data to training the network using Python, NumPy, PyTorch and evaluating performance through specialized metrics like the Dice Score.