You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This project uses PyTorch to classify bone fractures. As well as fine-tuning some famous CNN architectures (like VGG 19, MobileNetV3, RegNet,...), we designed our own architecture. Additionally, we used Transformer architectures (such as Vision Transformer and Swin Transformer). This dataset is Bone Fracture Multi-Region X-ray, available on Kaggle.
By comparing the performance of 24 custom trained models with various configurations, I aim to identify the most effective model architecture and configuration for accurate gel electrophoresis image classification.
Processing image advertisements to predict the context conveyed through them using CNNs. The images are further visualised using GradCAM to understand how the first and the last layers perceive the image dataset for the classification.
How AI enhances parking management. In a team of four, we leveraged deep CNNs to create a system that detects (YOLO) and classifies (ensemble of SqueezeNet and ResNet) parking spots as free or occupied from images obtained by common surveillance cameras. It is able to employ pre-installed cameras, optimizing space utilization and reducing costs.