Stanford’s Octopi Automated Imaging Platform Can Scan for Malaria in the Field

Scientists at Stanford University’s Prakash Lab have designed an automated imaging platform that can scan for malaria, tuberculosis, and…

Cabe Atwell
5 years ago

Researchers at Stanford University’s Prakash Lab have designed an automated imaging platform that can scan for malaria, tuberculosis, and other deadly diseases in the field, using readily available off-the-shelf parts. Octopi is a low-cost reconfigurable autonomous microscopy platform with automated slide scanning and correlated bright field and fluorescence imaging. The platform uses real-time computer vision and machine learning algorithms to screen more than 1.5 million red blood cells per minute to look for infection.

The Octopi can be outfitted with several different camera/imager modules depending on the application — such as a Raspberry Pi Camera and industrial imagers, which are used with a pair of cellphone or M12 focal lenses with autofocus capabilities to zoom in on blood cells depending on resolution needs. The same goes for the platform’s illumination modules, which range from using an LED panel (with diffuser and condenser) to lasers scavenged from DVD/Blu-Ray/CD players (for fluorescence illumination).

As with the camera/imager and illumination modules, the Octopi can also be outfitted with different control and computation modules that handle automated scanning and data processing, which includes the Raspberry Pi 3 Model B+, Nvidia Jetson Nano, Google Coral, and Intel NUC, each offering increased processing speed. Rather than buying a commercial motorized stage (found in high-end microscopes) for scanning applications, the scientists designed their own using a CNC machine aluminum block and a lead screw linear actuator that costs just $5.

The Octopi can be powered via either wall outlet, or a 5V DC battery pack, which the scientists found that a 20,000mAh battery could power the platform for more than eight hours on a single charge. As mentioned earlier, machine learning and computer vision are used to detect viruses and parasites, and the scientists were able to successfully identify Trypanosoma brucei rhodesiense (African sleeping sickness), Streptococcus pneumonia (community-driven pneumonia), and Staphylococcus aureus (respiratory infection).

Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles