Healthcare @ IBM Research | Australia - Multimedia Analytics


Multimedia Analytics @ IBM Research | Australia

Our researchers have expertise in video and image processing, data mining, pattern recognition and information retrieval. We feature the following projects to demonstrate our multi-media analytics capabilities.

 

Medical Sieve Grand Challenge

The Medical Sieve is a diagnostic clinical descision-support tool with advanced textual and visual reasoning capabilities.  As an image-guided informatics system, the Medical Seive acts like a filter, identifying and collating the clinical information required by physicians for patient diagnosis and treatment planning.  This filtering employs sophisticated medical image processing, pattern recognition and machine learning techniques guided by advanced clinical knowledge. It uses feature extraction and cognitive methods to automatically classify image content and retrieve similar patient cases based on image features.

Medical Sieve

Team members:  Simon Wail, Rahil Garnavi, Naveed Hashmi Suman Sedai, Xi Liang, Sisi Liang and Pallab Roy.

CollaboratorsIBM Research - Almaden and IBM Research - Haifa

 

Computer-Aided Melanoma Diagnosis

Melanoma (skin cancer) is one of the most common and deadly cancers world-wide, and has steadily increased in incidence over the last few decades. Diagnosis is a very subjective process, and consequently there is a lack of standardisation between institutions, between clinicians at the same institutions and even for different patients who see the same clinicain. This leads to a poor quality of treatment in areas where expertise are limited. IBM Research is collaborating with Memorial Sloan-Kettering Cancer Centre to  develop a system enabling standardised, early detection of melanoma. We automate the classification of skin lesions by application of image processing and machine learning algorithms to multi-modal skin imaging data, in combination with global, local and patient-level information, a-priori disease information, and temporal analytics.

Melanoma Image Processing
Modified. Original Image Paul Irish 

Team members: Rahil Garnavi, Mani Abedini, Rajib Chakravorty and Stefan von Cavallar

Collaborators: IBM T. J. Watson Research Center and Memorial Sloan-Kettering Cancer Centre

 


Selected Publications 

M. Abedini, N. Codella, J. Connell, R. Garnavi, M. Merler, S. Pankanti, J. R. Smith, and T. Syeda-Mahmood. A generalized framework for medical image classification and recognition IBM Journal of Research and Development vol. 59, no. 2/3 2015 (in press)

Q. Chen, M. Abedini, R. Garnavi, X. Liang IBM Research Australia at LifeCLEF2014: Plant Identification Task,  LifeCLEF2014 In proceeding of CLEF2014.

M. Abedini, L.Cao, N. Codella, J. H. Connell, R. Garnavi, A. Geva, M. Merler, Q. Nguyen, S. U. Pankanti, J. R. Smith, X. Sun, and A. Tzadok IBM Research at ImageCLEF 2013 Medical Tasks  American Medical Informatics Association, 2013

More publications.