Automated Malaria Parasite Detection

• A system for malaria parasite detection in thin blood-smears is presented. Sample slides prepared with standard methods are accepted. A low-cost TV-camera mounted on an ordinary microscope with a computer controlled stage is used as a picture sensor. Frames, digitized in windows of 64 x 64 pixels are fed into a special purpose picture processor at normal frame rate (25 frames/sec). In the picture processor measurements are made on the images at high speed. The classification problem is split into different levels each having different characteristics such as different sampling density. Four classes, three of which are different types of malaria parasites, are recognized. As a whole the classification is best labelled as a sequential pattern recognition procedure.In its preliminary version the system has been run at a speed comparable to that of a human operator, that is l 500 cells per minute. A test on 80 000 cells gave 25 false negatives out of 283 parasites (9%) and 41 false positives (0.05%).

Subject headings

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Signalbehandling (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Signal Processing (hsv//eng)

NATURVETENSKAP  -- Data- och informationsvetenskap -- Datorseende och robotik (hsv//swe)

NATURAL SCIENCES  -- Computer and Information Sciences -- Computer Vision and Robotics (hsv//eng)

TEKNIK OCH TEKNOLOGIER  -- Medicinteknik -- Medicinsk laboratorie- och mätteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Medical Engineering -- Medical Laboratory and Measurements Technologies (hsv//eng)

TEKNIK OCH TEKNOLOGIER  -- Medicinteknik -- Annan medicinteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Medical Engineering -- Other Medical Engineering (hsv//eng)

TEKNIK OCH TEKNOLOGIER  -- Medicinteknik -- Medicinsk apparatteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Medical Engineering -- Medical Equipment Engineering (hsv//eng)

Publication and Content Type

vet (subject category)

rap (subject category)