Human Generated Data

Title

Study of Three Warriors

Date

16th century

People

Artist: Adam Elsheimer, German 1578 - 1610

Classification

Drawings

Credit Line

Harvard Art Museums/Fogg Museum, Bequest of Meta and Paul J. Sachs, 1965.342

Human Generated Data

Title

Study of Three Warriors

People

Artist: Adam Elsheimer, German 1578 - 1610

Date

16th century

Classification

Drawings

Machine Generated Data

Tags

Amazon
created on 2020-05-02

Human 95.6
Art 93.1
Drawing 86.8
Person 80.1
Painting 76
Military 63.4
People 61.2
Military Uniform 59
Sketch 57.6

Clarifai
created on 2020-05-02

people 99.9
adult 99.1
print 98.9
art 98.8
man 98.8
illustration 98.7
wear 96.6
engraving 95
portrait 94.5
veil 94.3
woodcut 93.6
leader 92.3
group 91.7
chalk out 91.4
one 91.2
outfit 88.4
vintage 88.3
retro 87.7
etching 86.4
two 84.7

Imagga
created on 2020-05-02

sketch 100
drawing 86.4
representation 63.4
graffito 58.5
decoration 43.7
old 20.9
art 20.8
vintage 19
design 18.6
grunge 17
retro 16.4
black 13.8
symbol 12.8
stamp 12.3
detail 12.1
ancient 11.2
wall 11.1
letter 11
aged 10.9
postmark 10.8
postage 10.8
man 10.7
mail 10.5
antique 10.4
house 10
people 9.5
window 9.4
silhouette 9.1
sport 9.1
envelope 9
postal 8.8
textured 8.8
post 8.6
culture 8.5
winter 8.5
card 8.5
decorative 8.3
frame 8.3
pattern 8.2
paint 8.1
painting 8.1
paper 7.8
color 7.8
texture 7.6
competition 7.3
graphic 7.3
cartoon 7.1
architecture 7

Google
created on 2020-05-02

Microsoft
created on 2020-05-02

drawing 99.7
sketch 99.6
window 98.1
cartoon 95.3
text 93.7
art 81.7
old 80.8
illustration 79.5
black and white 75.5
painting 72.8
person 69.5
clothing 63.8
posing 37.8

Face analysis

Amazon

AWS Rekognition

Age 23-35
Gender Male, 54.9%
Surprised 47.3%
Angry 45%
Happy 45%
Calm 52.3%
Fear 45.1%
Disgusted 45%
Sad 45%
Confused 45.3%

AWS Rekognition

Age 15-27
Gender Male, 50.3%
Fear 49.5%
Angry 49.5%
Disgusted 49.5%
Happy 49.5%
Sad 49.7%
Surprised 49.5%
Confused 49.5%
Calm 50.3%

Feature analysis

Amazon

Person 80.1%

Captions

Microsoft

a vintage photo of a person 86.3%
a vintage photo of a person 84.9%
an old photo of a person 84.8%