Heart Rate Variability Analysis Application (HRV Explorer) for Smart Phone
Team: Prof. Dr. Wajid Aziz, Gulraiz Iqbal
Description: The cardiovascular diseases are major cause of mortality across the globe contributing 31% of global death burden. The electrocardiography (ECG) is used to detect abnormalities in the cardiac rhythms during the onset of cardiovascular problems, when damage has been already been done. The advances in smartphones and communication network technologies have revolutionized the modern world. Recently, smartphone enabled ECG monitoring systems have been developed, which record the ECG, display it on the screen, and transmit it to the physician/server for analysis. The device provides the information about mean heart rate (HR) and further analysis is made by the physician by visual inspection of the ECG. It is not possible for physician to visually inspect long term ECG recordings. It has been observed heart rate variability (HRV) is better predictor of cardiac abnormalities than HR itself. The reduced HRV shows changes in the cardiac autonomic control before it is manifested by HR on the disease onset. The incorporation of HRV Explorer in smartphone can assist physician to analyze long term ECG recording automatically for clinical decision making. Furthermore, patients or persons at risk of cardiac disease can use HRV Explorer which will provide summarized information of mean HR and various HRV indices.
It is widely accepted that technology can improve health of populations around the globe. The importance of smart phone technology in healthcare is at the forefront of the innovation. The smartphone is affordable and practicable for the both developed and developing countries. The proposed HRV Explorer can be use not only by physicians but also by patients for personal monitoring at home.
Prototype for Remote Infrastructures Surveillance
Funding: KACST-Lockheed Martin
Team: Dr. Ishtiaq Rasool Khan, Dr. Saleh Alshomrani, Dr. Syed Ahsan, Dr. Ali Hassan, Gulraiz Iqbal
Description: This is a KACST – Lockheed Martin funded research project. In this work, we are developing a prototype for remote surveillance, with aim to reduce the reliability of the surveillance systems on the human operators. The system consists of sensors of two different types – the proximity sensors to detect presence of a target, and video cameras to capture the scene. Concepts to be implemented and tested with this prototype include (a) development of data acquisition system, (b) power saving by intelligently activating the minimum required number of sensors, (c) bandwidth saving by discarding the irrelevant data before transmission, (d) development of video processing algorithms for real-time identification, and (e) using semantic and expert knowledge to assess the level of threat. Datasets will be collected under different situations and published publically, for purpose of development and evaluation of new algorithms
Virtual look-alike 3D Avatar for Health and Personality Enhancement
Team: Dr. Ishtiaq Rasool Khan (PI), Dr. Syed Ahsan, Gulraiz Iqbal, Mahboob Ali
Description: In this King Abdullah City of Science and Technology (KACST) funded project, we are developing an application which utilizes virtual reality to facilitate and motivate individuals and groups to improve their health. Core of the proposed research is to create 3D virtual avatars that look like the user or have any other desired body size and shape. Creation of customized avatars has been a challenging task that needs a lot of manual tweaking besides using expensive software tools and powerful hardware. In this proposed research, we will develop methods to create a realistic user- lookalike avatar (or one of any other shape and size) in a simpler way. To create a user-look-alike avatar, we assume the key body measurements of the user are available (taken manually or obtained by a scanning system). Two major modules that will be developed under this research are 3D mesh deformation, collision handling (to avoid penetration of body parts when size is modified) and 3D visualization and rendering. 3D interactive visualization tools will be developed using OpenGL, WebGL and C++. GPU will be used to achieve real-time performance.
Automatic Body Size Measurement and Finding Suitable Generic Clothing Sizes for Local Population
Team: Dr. Ishtiaq Rasool Khan (PI), Dr. Usman Saeed
Description:Database of body measurements of a different demographic populations are often created to carryout statistical analyses, to optimize the products and services for the population in a particular region. Collection of such measurements is a time-consuming process both for the subject whose measurements are collected, as well as for the expert collecting these measurements. Probability of human error is also high in manual measurements. On the other hand the available automatic techniques have their own issues. This research will address some of those and develop an automatic and efficient method of obtaining key body measurements using color and depth images of the user. New image processing and noise removal techniques will be developed to enhance the depth images. Features from both color and depth images will be integrated to get a segmented body profile of the user with accurate depth information of each pixel. Depth information will then be transformed to real world measurements using camera parameters which will be determined through a calibration process. The technology has applications in virtual training, education, medicine, and architectural designs. In local context, it can be used by the retailers in the Kingdom to provide right size of clothes to the customers coming from around the world, especially in Hajj and Umrah season. We will leverage on our technology to develop an interesting application to evaluate the generic clothing sizes of different retailers in the Kingdom. This can help them modify their designs to better serve the local population, and lead to better customer satisfaction as well as reduced waste and environmental fingerprint.
Encoding High Dynamic Range Images and Design of New Tone-Mapping Operators Suitable for Backward-Compatibility
Team: Dr. Ishtiaq Rasool Khan (PI), Dr. Muhammad Murtaza Khan
Improvement of HDR Coding Efficiency Through non-linear Quantization and Inverse tone-mapping Techniques
Funding: King Abdulaziz University, General Program
Team: Dr. Ishtiaq Rasool Khan (PI)
Suspicious Human Activity Recognition
Funding: University of Jeddah, General Program
Team: Dr. Usman Saeed (PI), Dr. Ishtiaq Rasool Khan
Description: In the last two decades video cameras have become omnipresent, from libraries, shops, train stations, airports and even homes. However the task of monitoring this 24/7 video feed is still inadequate as it depends largely on humans. The objective of automated human activity recognition is to automatically monitor the human activity alone and in groups and their interactions with objects/facilities to detect any abnormal events. Although such systems have not been deployed extensively but experimental systems have been made possible by ever increasing computational power and reduced cost.In this project we intend to develop a video surveillance system to automatically detect suspicious human activity. Image processing techniques will be developed for acquisition, representation, segmentation, tracking and recognition of human activities. Furthermore we shall also study the human behaviors and develop scenarios for normal/abnormal behavior. Several research opportunities exist in this domain , which include dealing with large open and closed spaces which require a multi camera approach, large crowds with leads to occlusion, daily and seasonal changes in crowd size and facility usage.
Dr. Ishtiaq Rasool Khan
University of Jeddah
Faculty of Computing and Information Technology
Usfan Road, Jeddah
Email : email@example.com
Phone : (012) 6952000 Ext. 74208