Because we view faces holistically rather than by individual features (nose, eyes, lips, etc.), some psychologists believe that current composite software is flawed. FACES, a commonly used composite software, is no exception to this. It gives witnesses a number of choices in terms of hair, face shapes, noses, and lips, but no choices to view a face as a whole.
A new approach, referred to as the whole-face approach, is slowly but surely being developed. Dr. Frowd is developing a system, called EvoFIT, that gives witnesses 72 choices (18 at a time), and are asked to pick the 6 that best resemble the perpetrator. When the 6 have been chosen, the software fuses the faces together which creates 72 more faces and the witness is asked to pick the best six again. The witness is then asked to pick the one face that best resembles the perpetrator as he or she remembers them and this images is used as the final composite.
Dr. Solomon is working on a similar, but slightly different software that supports the holistic approach.
Though these new software systems are currently being tested, the EvoFIT has been used in a sexual assault case where the composite apparently came so close to the perpetrator that it "caused the victim distress" (Dr. Frowd). It will be interesting to see the effectiveness of these new software programs down the road when they are used more frequently.
Below is a link to the article with more information about this new approach and a little about why current software is not the most accurate.
http://www.post-gazette.com/pg/07084/772371-84.stm
A new approach, referred to as the whole-face approach, is slowly but surely being developed. Dr. Frowd is developing a system, called EvoFIT, that gives witnesses 72 choices (18 at a time), and are asked to pick the 6 that best resemble the perpetrator. When the 6 have been chosen, the software fuses the faces together which creates 72 more faces and the witness is asked to pick the best six again. The witness is then asked to pick the one face that best resembles the perpetrator as he or she remembers them and this images is used as the final composite.
Dr. Solomon is working on a similar, but slightly different software that supports the holistic approach.
Though these new software systems are currently being tested, the EvoFIT has been used in a sexual assault case where the composite apparently came so close to the perpetrator that it "caused the victim distress" (Dr. Frowd). It will be interesting to see the effectiveness of these new software programs down the road when they are used more frequently.
Below is a link to the article with more information about this new approach and a little about why current software is not the most accurate.
http://www.post-gazette.com/pg/07084/772371-84.stm
I really enjoyed reading this article. It gave me a lot more perspective on facial composites.
I have never been a fan of the facial composites because I see and hear too many negatives on them, like going after the wrong person based on a poorly developed composite. In this article the composite drawn about the guy who sexually assaulted women in the East End was terribly wrong. It looked like two completely different men, and if the police were to start looking for a suspect with those facial features they would most likely looked over the real defendant.
I found it really interesting that they are switching to the holistic side of things, and I think they will find it to be much more helpful. I actually took the time to try and describe my boyfriend to a friend by using distinctive features, like eyes, nose, mouth, ears...and I found it to be difficult even though I have been around him and with him for 2 years. The one think I could do is see his face as a whole and I could actually think of people he looks like as a whole face. I agree 100% with the psychologists behind the holistic view.
All in all, this article was actually helpful for me, and very interesting as well!!
This is a great article! I didn't know that they made advances on the FACES system and did a little further reading on the EvoFIT from what appears to be the source. Below is the link to more information on the EvoFIT, how it was conceptualized, and so fourth. What I couldn't find in the readings is if this has been used yet for convictions. Also, my concern for a system in which you pick an entire face is if the witness will be "lead" to stick with an image closest to the first selection of the 18 given to the witness moreso than the actual appearance of the perpetrator. In short, my question is if this system could accidentally give the power of suggestion that the original choice is the "right" choice. Although this is definitely a step in the right direction, I would believe that a combination of the two systems may give a stronger lead to the perpetrator. For example; if the witness KNOWS eye shape and color, but the rest of the face is somewhat "a blurry memory"... we could start with selecting the eyes, then use EvoFIT for the rest.
http://www.evofit.co.uk/
I think that this article is great. I didn't realize that that was how the 'tested' to see whether the suspect was the actual criminal. I think it would be really tricky to be a victim and have to determine the true criminal because sometimes the crime was so fast and vapid that the victim may not remember all the details. So, EvoFIT is a great invention because it can decipher the criminal from someone who just looks like the criminal. I hope that by using this method there will be less and less innocent people being convicted of a crime.
I think that this article is great. I didn't realize that that was how the 'tested' to see whether the suspect was the actual criminal. I think it would be really tricky to be a victim and have to determine the true criminal because sometimes the crime was so fast and vapid that the victim may not remember all the details. So, EvoFIT is a great invention because it can decipher the criminal from someone who just looks like the criminal. I hope that by using this method there will be less and less innocent people being convicted of a crime.
I agree with the thought that EvoFIT will indeed help investigators, to some extent, in figuring out if the suspect is actually the perp. However, people have always had issues concerning memory of crimes and importantly how the suspect appeared. I feel that there will always be misconceptions associated with witness testimony. The witness or victim may only have a few very short seconds to memorize what happened during the event, let alone holistically remember a face.