The invention provides a face image
library construction method. According to the method, at least two different face images can be stored for the same object identified in a
library, and the face image
library can be used as a target face image library, a
video image face
image extraction library and a
video image face image summary library. The above method can be extended to the construction ofimage libraries of fingerprints, irises, palm prints, gaits, human heads, human bodies, car faces, whole cars, motor vehicles, non-motor vehicles, tires, unmanned aerial vehicles, aircraft, ships,
engineering machinery, tools, firearms, ammunition, cutters, bows, arrows, watches, mobile phones, texts,
graphics, nameplates, trademarks, clothes, trousers, belts, shoes, hats, jewelry, bags, flowers,leaves, fruits, seeds and insects, and can also be extended to the construction of the image libraries of dog and cat faces. According to the method, the missing report rate of identification when the face image library is used as the target library can be reduced, and the effective rate of images when the face image library is used as the extraction library or the summary library can be improved.