Welcome to my personal page
I received my BSc. degree in Information and Computer Science from Hadramout University of Science and Technology (HUST), Yemen, in 2005. Then, I completed my MSc. and Ph.D. degrees in Computer Science & Engineering at King Fahd University of Petroleum and Minerals(KFUPM), Saudi Arabia, in 2013, 2018 respectively. My research activities include pattern recognition, machine learning, and image processing (in general). Practically, I am interested in biometrics-based authentication systems, document image understanding, and biomedical engineering. In biometrics-based authentication systems, my research focuses on investigating multimodal authentication approaches using multiple physiological human treats such as facial, palmprint, and iris features. More details can be found here. In document-image understanding, it is noticed that several digital libraries worldwide lack fast document retrieval due to the presence of document-content in image format. Usually, these images are non-searchable by content and may delay information retrieval. Therefore, document-image understanding (DIU) plays an important role to preprocess these images into regions such as text, tables, figures, drawings etc. In addition, it determines the reading order/flow of these regions for final digitization of such content into searchable information such as text. In biomedical engineering, we focus on computational biology where we develop methods to detect abnormalities within medical images (e.g. MRI) or signals (e.g. ECG) that lead to disease diagnoses.
Manuscript Layout Analysis:
Usually, digital Arabic historical manuscripts are found in non-searchable forms such as images or scanned-PDFs. These forms delay how information can be searched or retrieved. A reader/researcher finds it a time-consuming process to retrieve a piece of information in a huge collection of manuscripts.
Because without text-based research a reader must find a
suitbale manuscript by title. Then, he/she is going to read
through its pages to find needed information.
Therefore, software-based solutions such as document layout analysis (DLA) and information retrieval (IR) can be used to allow searching manuscript images via examples (sub-images) or strings (text). In our research, we developed an accurate method to detect main-content of manuscript pages. Then, these main-contents are used for fast information retrieval and document classification using keyword spotting methods.
The electrocardiogram (ECG) is an important biomedical signal that is used to understand heart activity. For example, the exercise stress ECG is used as an indirect method for assessing individuals for limitation of coronary blood flow due to obstructive coronary artery disease(CAD). So, such limitation can be detected via analyzing the patient’s ECG signals. Our Last research focused on a method to integrate a 12-lead ECG system to enhance the detection of a limitation of coronary blood flow. read more..
Biometric Recognition and Authentication:
Personal recognition based on biometric technology is an emerging feild of research in security systems. Many researchers are attracted to this technology due to two main advantages; 1) It provides a good resistance to spoof attacks, and 2) It is hard to be lost or stolen. Typically, this technology uses human physiological or behavioral characteristics such as face, iris, palmprint, fingerprint, voice, gait, etc. for personal authentication or recogntion. Fortunately, each human-biotrait enjoys some strengths and suffers some weakness. In our research, we tried to utilize two or three individual human-traits such as face, iris and palmprint to develope better biometric authentication system. read more ...