DINOv3-Mask2Former: Urban Instance Segmentation
Unified instance segmentation framework combining self-supervised DINOv3 Vision Transformer with Mask2Former for urban streetscapes
Projects
Every project on this site: 25 of them, from instance segmentation for urban streetscapes to a firefighting robot and a CNC machine built from scratch. Cards with a title link lead to a full write-up.
All projects
Fast-Rich: High-Performance Terminal UI
High-performance Rust library for rendering rich text and complex terminal user interfaces
AraCLIP: Cross-Lingual Arabic Image Retrieval
Cross-lingual learning model for effective Arabic text-to-image retrieval using knowledge distillation from CLIP
Mapillary Vistas Dataset Visualizer
Interactive visualizer for the Mapillary Vistas semantic segmentation dataset
Structure from Motion Algorithm
Implementation of Structure from Motion algorithm for 3D reconstruction from multiple images
Text to Speech for Arabic Poetry
Text-to-speech system specialized for Arabic poetry with proper pronunciation and rhythm
MASC: Massive Arabic Speech Corpus
Large-scale Arabic speech corpus for automatic speech recognition research
RASA Chatbot with ChatGPT Integration
Integration of RASA framework with ChatGPT for enhanced conversational AI
RGB Arabic Alphabets Sign Language Dataset
Comprehensive RGB dataset for Arabic Alphabets Sign Language recognition
Arabic Sign Language Recognition on Raspberry Pi
Real-time Arabic sign language recognition system deployed on Raspberry Pi
Arabic Sign Language Recognition using Deep Learning
Deep learning models for Arabic Sign Language recognition
GANs for MNIST Digit Generation
Generative Adversarial Networks implementation for MNIST digit generation
MNIST Prediction using Neural Network
Neural network for MNIST handwritten digit classification
Sentiment Analysis using RNN
Recurrent Neural Network for sentiment analysis of text data
Word2Vec Embeddings Implementation
Implementation of Word2Vec algorithm using skip-gram architecture
