muhammadha04's repositories
VisionAssist_3D
VisionAssist is an augmented reality app for HoloLens 2, designed to improve navigation and object recognition for individuals with low vision. Leveraging Azure Object Anchors, it provides real-time 3D overlays to enhance spatial awareness and independence.
Detectron-2-Transfer-Learning
used the Detectron2 library and transfer learning to train a model for detecting persons, cars, and traffic lights in a custom dataset of street images taken in the Haifa area around the University campus.
HoloVisionAssist
HoloVisionAssist (MR app) leverages YoloV7 for real-time object detection via HoloLens 2, pinpointing and marking locations in 3D space with virtual pins.
JavaEats-Restaurant-Information-System
JavaEats is an information system with GUI designed to manage the daily operations of a restaurant that opened in Haifa during the Corona period.
WageGapTrends-LabProject
Exploring the Evolution of Gender Wage Gap Trends: The Impact of Health, Education, Demographics, and Labor Market Factors
Deep-Learning-and-Dropout-Based-Post-Training-Confidence-with-TensorFlow
the exercise also includes modifying the network architecture and activation functions to achieve better performance. Confidence estimation is also explored through the use of post-training dropout to generate multiple classifiers, with the confidence of the classification determined by the percentage of classifiers that voted for the same result.
Mask_RCNN
Mask R-CNN for object detection and instance segmentation on Keras and TensorFlow. Improved for 3D instance segmentation and additional evaluation options.
detectron2
Detectron2 is FAIR's next-generation platform for object detection and segmentation.
pytorch3d
PyTorch3D is FAIR's library of reusable components for deep learning with 3D data