In the process of making a dialysis filter, a number of fibers may be blocked by clear adhesive, and it is not possible for the naked eye to detect these blocked fibers, which are about 180 microns in diameter. These closed fibers accelerate the blood clotting process and reduce the efficiency of the dialysis procedure because fewer fibers are available for blood exchange. Coagulation can cause the volume of blood to be lost even when the extracorporeal circuit is completely clotted, so blood can no longer be returned to the patient. Therefore, it is necessary to identify the filters with closed fibers and remove them from the production line.
Fiber Detection Algorithm
For detection of close and open fibers in this system, we use deep learning networks.
Interface software
• Display processed images of 30 palettes
• Each palette contains 4 images of two filters
• Communicate with the tags of each palette via RFID
• Ability to save processed images automatically
• Display processed results including number of defects and processing time
