Prior to EFIT6, commercial, computerised techniques for constructing facial composites were essentially electronic versions of the original, mechanical feature-based systems such as PhotoFIT and Identikit. This approach, in which the face is compiled feature by feature, is fundamentally limited by the witness’s ability to recall from memory and verbalise accurate descriptions of facial features. A considerable body of evidence, assembled over several decades, now shows the human face perception mechanism is designed primarily to recognise faces not to recall and describe features. EFIT6 directly exploits the cognitively less demanding process of recognition.
In EFIT6, eyewitnesses are presented with groups of computer generated faces, broadly matching the initial description provided. By identifying both the strongest and weaker examples in each group, the user enables EFIT6 to learn and progressively present faces of increasing similarity to the suspect. A variety of sophisticated tools are also available for the witness to intervene and accelerate this process. The result is a system that is faster, easier and more accurate than previous composite technology.