Pennsylvania State University researchers are using image recognition technology to develop an automated method of classifying histopathological images.
"The idea is if you have a collection of images, can you automatically put them in categories?" asks professor Vishal Monga. He notes that of particular interest was that veterinarians with the university's Animal Diagnostic Laboratory (ADL) were capturing histopathological images of tissues. ADL's five pathologists examine more than 10,000 slides over the course of a year, says ADL veterinary pathologist Art Hattel. The time to adequately evaluate a slide can take between seven to 25 minutes, according to Hattel.
Graduate student Umamahesh Srinivas says Penn State electrical engineers designed tools to mimic how human pathologists classify tissue samples. Once these tools were developed, researchers achieved 80 percent to 85 percent accuracy in automatically categorizing the tissue samples in three areas: Healthy, inflammation, and necrosis, Monga says. He says the software could undergo a training phase to learn what healthy and diseased tissues looks like, after which it would be capable of processing thousands of images in a second. The researchers are now applying for larger grants to continue their effort to further develop the image recognition technology.
From Penn State Live
View Full Article
No entries found