Publications by authors named "Javairia Raza"
Article Synopsis
- The study focuses on using a convolutional neural network (CNN) to identify lipohypertrophy, a common skin issue from insulin therapy, in ultrasound images.
- A DenseNet CNN model achieved the best performance with 76% accuracy and recall in detecting this condition, with the use of image augmentation to enhance the dataset.
- The research also explored using the YOLOv5m model to pinpoint the exact location of lipohypertrophy in these images, showcasing the potential of machine learning in dermatological diagnostics.
