The output of A2IR2Lab is the first metric means by which it will be judged. Second, the
number of grants will contribute to the metric in an appropriate manner.
Researchers from A2IR2Lab shall aim to conduct works at the level that is publishable in
the most respected publication outlets of the field (conference proceedings and journal).
About
Research Programs
Computer-based Human Activity Understanding
Humanoid robots and sports coaching and surveillance cameras for public spaces are examples of the need for computer-based understanding of various classes of human activities. In this research program we focus on the development of computer vision-based algorithms for human activity detection, recognition, classication and understanding. Examples of projects in this program include:
• Continuous human activity recognition
• Segmented human activity recognition
• Human posture assessment
• Fine-grained human activity recognition
• Human interaction recognition
• Continuous human activity recognition
• Segmented human activity recognition
• Human posture assessment
• Fine-grained human activity recognition
• Human interaction recognition
AI for Food Security
As human population grows and the stress on arable land increases, there is need to address the attending food security issues. This research program focuses on the application of artificial intelligence through the design of computer vision, signal and image processing, and machine learning algorithms to solve food security problems. Example projects in this program include:
• Crop disease identification
• Crop disease database development
• Satellite- and drone-based land use classification
• Satellite- and drone-based vegetation identification and classification
AI for Health
This program focuses on the use of AI and Robotics techniques in solving computer-based diagnostics and analysis of various clinical imaging modalities. Examples of projects in this program include:
• Medical image registration
• Medical image segmentation (2D and 3D)
• Human gait classification and recognition
• Medical time-series signal analysis and classification
AI for Climate Change
One of the existential challenges faced by humanity is climate change. This manifests in tropical developing countries as desert encroachment, gully erosion (from abnormal rainfall pattern), prolonged dry seasons and attending farm harvest failure, air pollution, etc. Abatement, mitigation and rehabilitation strategies require building predictive models and monitoring the effect of implementations. These are all data and processing-hungry endeavours. Examples of projects in this program include:
• Land use monitoring, quantitative analysis and prediction
• Erosion monitoring quantitative analysis and prediction
• Air pollutants detection, monitoring and predictive modelling
• Contaminant detection, monitoring and quantitative analysis
• Hybrid power distribution modelling (micro-grids)
• etc.
Tools
This program focuses on developing deep understanding and building tools used in the general area of artificial intelligence, especially as it relates to the programs being pursued in the Lab. The projects in this program will of necessity be of deep mathematical, statistical avour mediated by engineering and real-life problems. Examples of projects in this program include:
• Matrix and tensor factorization for deep learning systems
• Information geometry
• Graph signal processing
• Sparse representation
• Harmonic analysis