Several demonstrations in LDV remind us that machines do not only learn through neural networks and machine learning. They have other ways to learn to recognize and analyze the world around them.
Research scientist Tali Dekel demonstrated a technology that uses computer machine vision to identify, which can be judged by magnifying deviations in straight lines or purple fruits on the roof.
On average, pathologists process 500 slides a day. There are thousands of individual cancer cells that need to be analyzed on each slide, and they are easily missed by doctors.
According to a study by the American Medical Association, usually less than half of pathologists agree with the correct diagnosis. Cited another example focusing on breast cancer lymph node biopsy, clarifying the difference between the research focus of computer and human pathologists. The former emphasizes many areas that may become containers for cancer cells.
The machine vision system provides pathologists with raw images, and then they can still view the data that they are familiar with and the images processed by the learning system. Basically, this can identify the area of cancer.