A lot has been recently reported and discussed about the eradication of malaria in Sri Lanka. While India certainly has a lot to learn from the neighbour's experience, the goal seems some way off, going by the spurt in vector-borne diseases across Indian cities during this time of the year, not least among them being malaria.
One of the principal factors behind reduction of malaria-related fatalities or complications is timely detection and treatment. Therefore, any technological breakthrough that can increase the speed and efficacy of diagnosis would greatly reduce mortality rates and ease the pressure on diagnostic clinics and hospitals.
Conventionally, the definitive diagnosis of malaria infection has been performed by a pathologist, where the presence of the parasite is manually examined in stained blood smear slides under a light microscope. The major drawback of such manual examination is subjectivity, which varies from pathologist to pathologist in terms of their experience and knowledge. Additionally, it is of course a time consuming evaluation process.
In order to provide a reliable and rapid malaria detection solution, KGP researchers have developed a smartphone-based android app for automatic recognition of parasite-infected red blood cells (RBCs) and its counting. This work was done in the Bio Medical Imaging Informatics (BMI) Lab, led by Chandan Chakraborty and his research team at the School of Medical Science & Technology, IIT Kharagpur in collaboration with Ashok Kumar Maity, pathologist, Midnapur Medical College Hospital. The research on "Medical imaging informatics for malaria detection" was funded by CSR grant from Microsoft Corp. Microsoft through their giving partner CAF India.
The smartphone is attached by an adapter with the eye-piece of the existing conventional light microscope, already available in any pathology lab. The stained blood smear slide is then kept under the microscope and the camera of the smartphone grabs many microscopic image frames from the slide. These grabbed images can even be saved digitally and then the app is used to detect all the infected RBCs present in the image.
The technology is expected to be particularly useful for pathology labs, where especially no expert pathologist is available. Para-medical or trained personnel would also be able to use it for accurate malaria detection, which holds promising implications for treatment of the disease in rural or remote areas. "The app uses microscopic image analytics and machine learning techniques for automatically detecting only malaria parasite-infected RBCs among other RBCs from microscopic image frames. It can be very useful for rapid malaria screening under tele-pathology framework," Chakraborty said.
Till now, the app has been tested on more than 200 microscopic images of 80 patients as a pilot study with more than 90% accuracy. As the "malaria diagnostic" app is developed in an open source platform, it may be of very nominal cost or no cost at all.
Source: http://timesofindia.indiatimes.com/city ... 935514.cms
Who is online
Users browsing this forum: No registered users and 0 guests