Researcher Thaer Darabi', a student in the master's program of data science and business analysis, defended his thesis entitled: "Using Competitive Generative Networks in Retrieving Images Based on Manual Graphics."
The thesis presented a research in the field of manual graphics-based image retrieval (SBIR) to find similar images from big databases by entering sketches into a model composed of a generative unit and a retrieval unit. This model aims to develop the ability to view and retrieve images from big databases, and to solve the problem of linking manual drawings to groups of images that are seen during training and that fail to generalize to groups that are not visible. The production unit (Generative adversarial networks) depends on tools to extract features and characteristics from images to get advanced results in image retrieval.
The thesis was supervised by Professor Muhammad Awad, and examined by Dr. Amani Owda and Dr. Muhammad Al-Dasht.