Since Paul Viola and Michael Jones, two researchers from 200 1 Microsoft, invented a frame algorithm for judging arbitrary graphics in images and applied it to face recognition, the algorithm of face recognition has been constantly evolving through years of improvement and evolution. Nowadays, face recognition technology is becoming more and more perfect in weak card cameras (compared with SLR cameras and smart phones). Although the face recognition algorithms of different manufacturers are different due to patent competition and other factors, their basic principles are not very similar.
. By processing the information such as color, gray level and brightness of the image, the solid color or variegated color parts are removed first, and then the graphics close to the eye, eyebrow, mouth, nose and other features are found, and the similarity is compared with the built-in face graphics, and the graphics whose similarity meets the conditions in the mathematical sense are taken out and judged as "faces".
. Because the comparison algorithms of different manufacturers are different at present, not only faces (including faces in pictures), but also pets and humanoid figures (such as the snowman in this case) will be identified as long as there are enough features similar to faces. It is worth mentioning that the smiling face shutter technology is very similar to face recognition, but the focus is more on the shape judgment of mouth and lips.