View
52
Download
0
Category
Preview:
DESCRIPTION
Computerized classification of large ceramic assemblages A quantitative basis for relative chronology. Avshalom Karasik The Computerized Archaeological Laboratory, Hebrew University of Jerusalem. Who are we?. Uzy Smilansky. Ilan Sharon. Leore Grosman. Talia Goldman. A vshalom Karasik. - PowerPoint PPT Presentation
Citation preview
Computerized classification of large ceramic assemblages
A quantitative basis for relative chronology
Avshalom Karasik
The Computerized Archaeological Laboratory, Hebrew University of Jerusalem
Who are we?
Uzy Smilansky
Leore Grosman
Talia Goldman Avshalom Karasik
Ilan Sharon
Ortal Haroch
"Pottery is, however, the greatest resource of the archaeologist. For variety of form and texture, for decoration, for rapid change, for its quick fall into oblivion, and for its incomparable abundance, it is in every respect the most important material for study, and it constitutes the essential alphabet of archaeology in every land.…to tie together sequences found at related sites in a region to form a master chronological sequence. This would enable any absolute dates determined from one site (for example through inscriptions, documentary evidence, and so on) to be transferred to other sites in the master sequence”
F. Petrie 1904
"Pottery is, however, the greatest resource of the archaeologist. For variety of form and texture, for decoration, for rapid change, for its quick fall into oblivion, and for its incomparable abundance, it is in every respect the most important material for study, and it constitutes the essential alphabet of archaeology in every land.…to tie together sequences found at related sites in a region to form a master chronological sequence. This would enable any absolute dates determined from one site (for example through inscriptions, documentary evidence, and so on) to be transferred to other sites in the master sequence”
F. Petrie 1904
A Test Case – Torpedo Storage Jars
The assemblage: Torpedo storage jars from Hazor (53) and Tyre (24), that have been already discussed in the literature in regards to their similarity and its historical consequences.
• Geva S. BASOR 248 Pp 69-72. 1982. Bikai P. BASOR 258 Pp 71-72. 1985. Gilboa A. In QEDEM Reports 2, page 11. 1995.
. א , - לאמר אלי יהוה דבר -בויהי , - על שא אדם בן . צרואתה , גקינה שבת הי לצור ואמרת , , , ; - , , כלילת- אני אמרת את צור יהוה אדני אמר ה כ רבים ים אי אל העמים רכלת ים מבואת על
, יזיפי. ... רכליך המה ישראל וארץ , יהודה , , ; נתנו וצרי ושמן ודבש ופנג מנית בחטי מערבך.
The word of the Lord came to me: 2Now you, mortal, raise a lamentation over Tyre, 3and say to Tyre, which sits at the entrance to the sea, merchant of the peoples on many coastlands, Thus says the Lord God: O Tyre, you have said, ‘I am perfect in beauty.’ … 17Judah and the land of Israel traded with you; they exchanged for your merchandise wheat from Minnith, millet,* honey, oil, and balm.
Ezekiel 27
A Test Case – Torpedo Storage Jars
A correlation matrix
0.4
0.5
0.6
0.7
0.8
0.9
1
10 20 30 40 50 60 70
10
20
30
40
50
60
70
Hazor
Tyre
Jar Index
Jar Index
A Test Case – Torpedo Storage Jars
214143 91018425348202529542669386162707727724965687445735556576359586466606771 1 4 51475 2 7223932 3362335407644 837503446 6473051131719312833111624521215-5
0
5
10
15
20
25
30
35
40
45
A Test Case – Torpedo Storage Jars
• Tyre• Hazor
• There is a significant morphological difference between the jars from Tyre and those from Hazor.
• Any theory that describes the commercial connections between the two sites based on the similarity of the jars must explain this independent dichotomy.
• For instance: it is possible that the differences are due to the fact that the two assemblages were drawn by two different draftspersons.
• The archaeological publications must progress to automatic digital documentation of ceramic.
A Test Case – Torpedo Storage Jars: Conclusions
High precision data acquisition using 3D scanners.
We have developed an innovative algorithm which automatically finds the symmetry axis of pottery fragments and extracts their profiles.
Final products – automatically produced print quality drawings
Final products – automatically produced print quality drawings with colored views
0 10 20 cm
0 70 1400
40
80
X
Y
-100 -50 0 50 100110
115
120
125
130
135
140
Arc-length
Rad
ius
-100 -50 0 50 1000
1
2
3
4
5
6
Arc-length
Tang
ent
-100 -50 0 50 100-0.2
-0.1
0
0.1
0.2
0.3
Arc-length
Cur
vatu
re
=
Automatic classification and typology of ceramic fragments
0 70 1400
40
80
X
Y
-100 -50 0 50 100100
110
120
130
140
Arc-length
Rad
ius
-100 -50 0 50 1000
1
2
3
4
5
6
Arc-length
Tang
ent
-100 -50 0 50 100-0.2
-0.1
0
0.1
0.2
0.3
Arc-length
Cur
vatu
re
Automatic classification and typology
0 70 1400
40
80
X
Y
-100 -50 0 50 100100
110
120
130
140
Arc-length
Rad
ius
-100 -50 0 50 1000
1
2
3
4
5
6
Arc-length
Tang
ent
-100 -50 0 50 100-0.2
-0.1
0
0.1
0.2
0.3
Arc-length
Cur
vatu
re
Automatic classification and typology
0 70 1400
40
80
X
Y
-100 -50 0 50 100110
115
120
125
130
135
140
Arc-length
Rad
ius
-100 -50 0 50 1000
1
2
3
4
5
6
Arc-length
Tang
ent
-100 -50 0 50 100-0.2
-0.1
0
0.1
0.2
0.3
Arc-length
Cur
vatu
re
Automatic classification and typology
8 10 12 14 16 18
1
2
3
4
5
6
Bowl
12 14 16 18 20 2218
19
20
21
22
23
24Krater
8 10 12 14 16 182
3
4
5
6
7
8Cooking pot
0 2 4 6 8 106
7
8
9
10
11
12Jug
0 2 4 6 8 10 1223
24
25
26
27
28
29Jar
-10 -8 -6 -4 -2 0 2 4 6 8 102
4
6
8
10
12
14
16
18
20
Arc-length
Rad
ius
BowlKraterCooking potJugJar
-10 -8 -6 -4 -2 0 2 4 6 8 100
1
2
3
4
5
6
Arc-length
Tang
ent
BowlKraterCooking potJugJar
-10 -8 -6 -4 -2 0 2 4 6 8 10-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Arc-length
Cur
vatu
re
BowlKraterCooking potJugJar
Automatic classification and typology
Fragment Index
Frag
men
t Ind
ex
100 200 300
100
200
300
0
0.5
1
1.5
2Distance Matrix
Automatic classification and typology
0
500
1000
1500
2000
2500
3000
3500
Jars and Jugs
Open bowls
variacraters
Cooking pots and Craters
Primary clustering: Example: 755 Iron-age fragments from area G in Tel-Dor (Israel)
0
500
1000
1500
2000
2500
3000
3500
0 10 20 cm
0 10 20 cm
0 10 20 cm
122123110
120
100 300
200
213110
220
212
221 222
214 211 311313 314 312121
321 322315
330
230
230 210330310 320
Detailed clustering: Example: 380 Iron-age fragments from area G in Tel-Dor (Israel), that were already analyzed by Ayelet Gilboa (Haifa University).
110 121 122 123 211 212 213 214 221 222 230 311 312 313 314 315 321 322 3300
10
20
30
40
50
60
70
Type-Code
Num
ber
of v
esse
ls
BowlsKratersCooking potsJugsJars
Detailed clustering: Example: 380 Iron-age fragments from area G in Tel-Dor (Israel), that were already analyzed by Ayelet Gilboa (Haifa University).
The 300 subtypes
0 10 20 cm
1 2 3 4 5 6 7 8 9 10
11 12 13 14 15 16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31 32 33 34 35 36 37 38 39
Type-Code 311 Type-Code 313Type-Code 312
Type-Code 314 Type-Code 315
Type-Code 321
Type-Code 330Type-Code 322
Detailed clustering: Example: 380 Iron-age fragments from area G in Tel-Dor (Israel), that were already analyzed by Ayelet Gilboa (Haifa University).
-50 -40 -30 -20 -10 0 10
-0.1
-0.05
0
0.05
0.1
Arc-length (mm)
k(s)
Type-code 211Type-code 212Type-code 213Type-code 214
type-code 211type-code 212
type-code 213 type-code 2140 10 20 cm
Discriminant Analysis of the results
Original type-codes
DA
labe
ling
100 200 300
100
200
300
0 10 20 cm
0 10 20 cm
110 120
120
110
310 320 330
330
320
310
Original type-codes
DA
labe
ling
210 220 230
230
220
210
Kefar Hananya : Late Hellenistic - early Roman pottery production center. Can one distinguish systematic style variations between different workshops in the same village?
High resolution typology
Bowls type 1E
Cooking pots type 4C
In collaboration with Prof. David Adan-Bayewich (Bar Ilan University)
E1 bowls
High resolution typology
0
20
40
60
80
100
Clu
ster
hei
ght
Main excavation areaSquare O4
1 2
1b 1a
3
Discriminant Analysis
0 0.5 10
10
20
30
40
50
60
70
Provenance Index
%
Square O4Main excavation area
Chemical analysis10 element abundances measured by neutron
activation and high-precision X-ray fluorescence analyses
Shape analysis
Summary and Conclusions
This presentation describes a novel method for computerized ceramic typology and classification which is based on a numeric distance function between three mathematical representations of the profile.
Automatic classification can follow the traditional archaeological concepts of typological classification and even improve its resolution.
The classification procedure is fast, objective, reproducible and has no bias from subjective judgments.
Computerized classification of large ceramic assemblages can and should serve as a quantitative basis for relative chronology.
Thank you for your attention!
122123110
120
100 300
200
213110
220
212
221 222
214 211 311313 314 312121
321 322315
330
230
230 210330310 320
110 121 122 123 211 212 213 214 221 222 230 311 312 313 314 315 321 322 3300
10
20
30
40
50
60
70
Type-Code
Am
ount
BowlsKratersCooking potsJugsJars
0 70 1400
40
80
X
Y
-100 -50 0 50 100100
110
120
130
140
Arc-length
Rad
ius
-100 -50 0 50 1000
1
2
3
4
5
6
Arc-length
Tang
ent
-100 -50 0 50 100-0.2
-0.1
0
0.1
0.2
0.3
Arc-lengthC
urva
ture
Recommended