Face recognition, the neural aspect

Tips, trick's, and links about computer graphics world

Moderator: joepal

Face recognition, the neural aspect

Postby joepal » Sat Jun 03, 2017 6:58 am

Probably very off topic, but might be good to know if you have an interest in faces and psychology.

Popular science summary: https://www.scientificamerican.com/arti ... tion-code/

Full article: http://www.cell.com/cell/fulltext/S0092-8674(17)30538-X?_returnURL=http%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS009286741730538X%3Fshowall%3Dtrue
Joel Palmius (LinkedIn)
MakeHuman Infrastructure Manager
http://www.palmius.com/joel
joepal
 
Posts: 4465
Joined: Wed Jun 04, 2008 11:20 am

Re: Face recognition, the neural aspect

Postby RobBaer » Sat Jun 03, 2017 1:53 pm

As a guy limited binocular capabilities because my eyes are fairly close together, it's interesting that this study is apparently based completely on 2D morphing rather than 3D patterns with depth: (http://www.cell.com/cms/attachment/2095 ... 5/mmc1.mp4) . Does this mean that binocular perceptions are of minor importance. My recollection is that facial recognition is on the inferior portions of the cortex rather than in the occipital lobe and surrounding visual association cortex. The article seems to focus on "recognition". I wonder if if our "emotional reads" involce processing by those same neurons.

When I look a the control points, though, I'm reminded of Manuel's choices for face rig bone positions. Because MH doses't "come with hairline", and hairline seems an intrinsic part of the recognition system, perhaps the facial rig will require some tweaking for 3D psychology work. A hair system or systems remains a limitation for MH at present, but bones could move poly hair much like the 2D morphing program. It will be interesting to see how MB Labs approaches this issue.if it ever goes there. I wonder if lattice-based rigging has any benefits over bone-based rigging for facial recognition and expression systems.
User avatar
RobBaer
 
Posts: 1208
Joined: Sat Jul 13, 2013 3:30 pm
Location: Kirksville, MO USA

Re: Face recognition, the neural aspect

Postby blindsaypatten » Sun Jun 04, 2017 3:32 pm

The fact that humans are so adept at identifying faces in photographs is good evidence that binocular vision isn't important to facial recognition. I highly doubt that facial recognition is significantly impaired by monocular vision either.

I think bones for hairline movement would be problematic as you need the hair to follow but not change the contours of the skull. I could see morphing the scalp mesh to move the hairline while maintaining the shape of the skull. I was able move eyebrows by moving the mesh lines they were attached to and then shrink wrapping back to the original surface shape, but the shapekey has to be specific to the particular skull shape as the offsets differ depending on the shape of the mesh being "slid over".
blindsaypatten
 
Posts: 586
Joined: Tue Mar 14, 2017 11:16 pm

Re: Face recognition, the neural aspect

Postby loki1950 » Tue Jun 13, 2017 4:12 am

Just listened to a rebroadcast from CBC our national public radio there 30yr old science show Quirks & Quarks http://www.cbc.ca/radio/quirks they had a section on this Saturday with the repeat tonight you all can listen from there web site http://www.cbc.ca/radio/quirks/your-bab ... oplay=true it includes a separate section on a functional MRI study on monkey facial recondition augmented by neural probes in key areas it appears that there are 50 dimensions that describe any face.

Enjoy the Choice :)
my box::HP Envy i5-6400 @2Q70GHzx4 8 Gb ram/1 Tb(Win10 64)/3 Tb Mint 19.2/GTX745 4Gb acer S243HL K222HQL
Q8200/Asus P5QDLX/8 Gb ram/WD 2Tb 2-500 G HD/GF GT640 2Gb Mint 17.3 64 bit Win 10 32 bit acer and Lenovo Ideapad 320-15ABR Win 10/Mint 19
User avatar
loki1950
 
Posts: 1219
Joined: Thu Dec 18, 2014 6:27 pm
Location: Ottawa,Ontario


Return to Off topic

Who is online

Users browsing this forum: No registered users and 1 guest