Showing papers on "Facial recognition system published in 1979"

Propriétés spectrales des algèbres de Banach

Abstract: and Preliminaries on Biometrics and Forensics Systems.- Data Representation and Analysis.- Improving Face Recognition Using Directional Faces.- Recent Advances in Iris Recognition: A Multiscale Approach.- Spread Transform Watermarking Using Complex Wavelets.- Protection of Fingerprint Data Using Watermarking.- Shoemark Recognition for Forensic Science: An Emerging Technology.- Techniques for Automatic Shoeprint Classification.- Automatic Shoeprint Image Retrieval Using Local Features.

On Training People to Recognize Faces

Abstract: Experiments were carried out in an attempt to validate the facial recognition element of a three-day training course in person recognition. In one experiment the subjects had to rely on memory. In two other experiments they were provided with an array of full-face photographs to match against different poses and expressions of the same faces. The performance of subjects who had received training was never reliably better than that of untrained subjects, and in one experiment was significantly worse. Ii is suggested that the emphasis on isolated facial features in the training course may be responsible for its lack of success, and that processing independent features is not a good basts for a model of facial recognition

Processing multiple physical features in facial recognition.

Abstract: Subjects made decisions about a series of faces, then had a recognition test The study tasks required a decision about self-reference (does it look like you), an abstract trait (eg, friendliness), a single physical feature (eg, thickness of lips), or multiple physical features (which is the person’s most distinctive feature) As in past research, single physical feature decisions led to poorer performance than abstract trait decisions Of greater interest here, the multiple physical feature condition was significantly better than the single physical feature condition, though still slightly worse than the abstract trait condition These results indicate that the number of features processed is as important to facial recognition as feature “depth,” and that under some circumstances encoding “shallow” attributes can produce durable memory traces