Automated Camera Focusing

Development Challenge #1 – Autofocus

One of the first hurdles encountered was automating the focus of the camera.  We had to replicate the actions of the human eye and the microscope’s focusing mechanisms to automate the image-capture process.  This is a critical step in the total automation process since the quality of the images captured has a direct correlation to the performance of the detection algorithms.  Several factors needed to be accounted for in order to maximize the autofocus process: (i) smears are topographically uneven – having no smooth or even surfaces; (ii) the glass slides may be unevenly cut resulting in z axis distortion; and, (iii) the slides may not set evenly on the microscope stage.  An additional challenge in the autofocus process is the requirement to collect a minimum of 100 fields-of-view (FOV) from each slide.  A technological advancement was required that would maximize focus across the specimen, without the need for refocusing before each FOV capture.  The issue with refocusing each FOV captured was the added time required, which would have significantly and negatively impacted throughput performance. Our technical team developed a unique technological solution to these challenges that is currently part of our patent portfolio, pending final U.S. Patent Office acceptance.

Development Challenge #2 – Auramine Stain

To the trained microscopist, Auramine stain has as one of its properties, the ability to bloat TB bacteria making it easier for detection.  Auramine is a fluorescing stain that causes the image background to range from black to fuzzy green with little visual contrast.  Attempting to autofocus the camera on the Auramine stained slides can best be equated to using your personal digital camera, in the fully automatic setting, and taking pictures of the sky on an overcast day.  It appears there is nothing for the camera to focus on.  Overcoming this challenge required many months of research, creation of many potential solutions, and the testing of each possible solution.  The result was the development of a patent-pending focus process.

Development Challenge #3 – Ziehl-Neelsen (ZN) Stain

ZN stain has higher levels of contrast, but is highly viscous which results in smears with high peaks and low valleys.  Because of those characteristics, TB bacilli can be found anywhere in the various depths of the smear.  Managing a focused image simply means that the image is in focus for those objects within that focal plane.  Objects located above or below the focused image plane are either blurred or not captured at all.  We have researched, developed, and tested another patent-pending process to capture in-focus images where all objects at the various depths appear in focus.

In the ZN staining process, counterstain is used to provide image contrast.  Sometimes the objects absorbing the ZN stain become masked by the counterstain.  To the human eye this creates clutter and confusion that obscures the detection of TB bacilli.  Once again, our technical team, after a tremendous amount of time and effort, has created a process that removes the counterstain mask.

Development Challenge #4 – Oil Immersion

Unlike Auramine stain, ZN does not bloat the bacteria to aid the human visual process.  To compensate for this lack of bloating effect, microscopist must use drops of oil on the ZN stained slides to magnify the bacilli and render a possible bacilli detection.  Use of oil with the hardware components listed previously would not be feasible in an automated process.  Therefore a non-oil solution had to be researched and developed.  Once again the technical team has been able to overcome this challenge through the development of a patent-pending process that eliminates the oil, yet provides visually detectable bacilli, even to the human eye.

Advertisements
This entry was posted in Corporate, Healthcare, TB and tagged , , , , , , , , , , , , . Bookmark the permalink.