Founder
Current• Implemented a data collection system, utilizing MediaPipe holistic keypoints, capturing diverse hand gestures and movements essential for model training.• Employed advanced data augmentation techniques to enrich Asher's dataset, ensuring robustness and diversity in training samples.• Developed a hybrid Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) architecture tailored for the complexities of ArSL recognition and interpretation.• Integrated a translation system featuring contextual phrase generation, enabling Asher to provide nuanced and contextually relevant translations of ArSL gestures into Arabic text.• Collaborated with local deaf organizations, to enhance Asher’s capabilities and expand its reach.• Partnered with local businesses, to tailor the translator for specific industry-related sign language needs.