Researchers at the Audio and Image Research Laboratory in the Department of Electronics under the Cochin University of Science and Technology have developed a software system called ‘SpineSearch’. The software also facilitates content-based search for similar images in the database. For the rectification of this spinal curve, a measure of the spine curvature (Cobb angle) is required.
New algorithm
Tessamma Thomas, Professor in the Department of Electronics, who guided the research work carried out by Dinesh Kumar P.V., told The Hindu on Monday that the new rule-based algorithm developed for strategic vertebrae selection and Cobb angle and other scoliosis feature evaluation measurements has been checked for its accuracy and reliability.
“At present, medical image retrieval with traditional Picture Archiving and Communication Systems (PACS) & Electronic Medical Records (EMR) systems is based on textual descriptions like filenames, keywords, patient identity number etc. stored with the image. But visual features help in retrieval of cases of patients having similar diagnosis and also cases with visually similar but different diagnoses,” she said.
Award
Winner of the Rashtriya Gaurav Award (2009) and the Shiksha Rattan Puraskar (2010) instituted by the India International Friendship Society, Dr. Thomas said that the Content Based Image Retrieval (CBIR) system (using text and visually derived feature information) helps in integrating scoliosis features and the treatment procedure adopted along with pre-operative and post-operative images.
“Thus, a surgeon can search for a similar previous case in the database and assess the outcomes of various techniques for a given case in hand, to decide the surgical strategy. This can be used in telemedicine application wherein a spine surgeon from a distant place can access these details, if they are available in databases which are accessible worldwide. Also, a trained user can also use the system for diagnostic purpose in large scale initial scoliosis screening,” she said.