Facial Image Features, Text Recognition and Segmentation
Dara O Shayda
dara@compclassnotes.com
dara@compclassnotes.com
In[]:=
DateObject[]
Out[]=
© 2012-Present CCN StudiosCreative Commons Attribution-NonCommercial-ShareAlike 4.0
Concept
● Image Identity blockchains with rich contents
● Arbitrary but small metadata, larger can be added to the contents with ease
● Batch processing
● Image Hash requirements same as the String type
● Augmented additional certificates 👥💬
● Arbitrary but small metadata, larger can be added to the contents with ease
● Batch processing
● Image Hash requirements same as the String type
● Augmented additional certificates 👥💬
Any Image Orientation
Out[]=
Extract all Faces
Out[]=
Align Faces
Out[]=
Facial Features
Out[]=
Outline Points: Face, eyes, nose and mouth left vs. right
Out[]=
outlinepoints{{5.85156,70.7188},{5.85156,59.0938},{10.8672,47.4688},{17.5547,35.8438},{20.8984,26.1563},{27.5859,18.4063},{32.6016,8.71875},{42.6328,4.84375},{54.3359,4.84375},{64.3672,6.78125},{74.3984,10.6563},{81.0859,18.4063},{86.1016,26.1563},{89.4453,35.8438},{89.4453,47.4688},{101.148,61.0313},{101.148,72.6563}},righteyebrowpoints{{10.8672,78.4688},{17.5547,86.2188},{27.5859,88.1563},{37.6172,86.2188},{44.3047,84.2813}},lefteyebrowpoints{{61.0234,84.2813},{69.3828,88.1563},{79.4141,88.1563},{87.7734,86.2188},{94.4609,80.4063}},nosepoints{{52.6641,74.5938},{52.6641,66.8438},{52.6641,57.1563},{54.3359,49.4063},{44.3047,43.5938},{49.3203,43.5938},{54.3359,41.6563},{59.3516,43.5938},{62.6953,43.5938}},righteyepoints{{24.2422,74.5938},{27.5859,78.4688},{34.2734,78.4688},{39.2891,72.6563},{34.2734,72.6563},{27.5859,72.6563}},lefteyepoints{{67.7109,74.5938},{72.7266,78.4688},{77.7422,76.5313},{82.7578,74.5938},{79.4141,72.6563},{72.7266,72.6563}},mouthexternalpoints{{35.9453,30.0313},{42.6328,31.9688},{49.3203,33.9063},{54.3359,31.9688},{59.3516,33.9063},{66.0391,31.9688},{72.7266,30.0313},{66.0391,24.2188},{59.3516,22.2813},{54.3359,22.2813},{49.3203,22.2813},{42.6328,24.2188}},mouthinternalpoints{{37.6172,30.0313},{49.3203,28.0938},{54.3359,28.0938},{59.3516,30.0313},{71.0547,30.0313},{59.3516,28.0938},{54.3359,28.0938},{49.3203,28.0938}}
Image Text Extraction
Out[]=
UNITED STATES OF AMERICĂPASSPORT CARDNAME DEANTEANCERIC WILLIAMSNATICNAIES DIMRC.MO0KUNITED STATES OF AMERICADATE NTE BIRTH23 DEC/DEC 1036EXPIRINE 6 RATEE/DEC 1988NEAISS NUNMERSA5SKON DEC 1986.2X18ta. C.26CPUSAERIC<<WILLIAMS<<<<<<WB1002XXRX<<<<<ZUA8612230M r92112046M2xxxxxxxxxxxxxxxxxxx
Identity Blockchains with more than one photo 👥💬
Mood Detection Dataset 👥💬
Out[]=
Image Measures to Cull Duplicates
Often Identity Cards have multiple different sized and different colorized images of the same person. There are image measurement functions available to cull the duplicates.
Out[]=
,
Foreign Identity Cards
Language “ChineseSimplified”, Method”Tesseract”