OIPL Statistics

Fri, 10 Jul 2026 12:11:40 -0500

Otolith Classification With Neural Embeddings and ICP

Otolith Classification With Neural Embeddings & ICP

Individual Statistics

This section presents statistics for individual otolith samples.

An otolith with a red outline and dark red background is misclassified (and none of its 9-nearest images have the correct label ➔ worst case).

An otolith with a red outline and pale red background is misclassified (its 1-nearest image has the wrong label, but at least one of its 9-nearest images has the correct label).

An otolith with a orange outline is well classified, but its 2-nearest image has the wrong label.

An otolith with a yellow-green outline is well classified, but its 3-nearest image has the wrong label.

Otherwise, the outline is green (its 3 first nearest-images have all the correct label ➔ ideal case).

105 classes: Achlin, Adixen, Anchep, Anclyo, Ancmit, Ancomm, Arcpro, Arifel, Bagmar, Baichr, Baichr2, Batsop, Brepat, Carcry, Carhip, Carhip2, Cenphi, Censtr, Chafab, Chasab, Chisch, Chlchr, Citmac, Citspi, Citspp, Cryros, Ctebol, Cynare, Cynneb, Cypvar, Dipfor, Dormac, Dorpet, Erosma, Eucarg, Eucgul, Euchar, Eucmel, Eucspp, Evolyr, Fun, Fungra, Funmaj, Funpul, Funsim, Gamhol, Gerreidae, Gobbos, Goboce, Gobstr, Harjag, Hypmee, Lagrho, Larfas, Leixan, Lucpar, Lutcam, Lutgri, Lutsyn, Memmar, Mename, Menber, Menlit, Menpen, Menspp, Micund, Mugcep, Mugcur, Mugspp, Mycmic, Myrpun, Nicust, Olisau, Opiogl, Opsbet, Ortchr, Paralb, Parlet, Pepbur, Peppar, Poelat, Pogcro, Pomsal, Priros, Pritri, Raccan, Scaridae, Scioce, Selvom, Sparad, Sparub, Sphbar, Sphgua, Sphpar, Stecap, Stelan, Stelan2, Strtim, Sygspp, Sympla, Synfoe, Synlou, Tracar, Trilep, Trimac.

Class #0: Achlin

Number of samples : 2

Classification Accuracy : 50% (1/2)

Most confused with : Sympla, Citmac, Leixan, Fungra, Lagrho, Stecap

Class #1: Adixen

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Fun, Cypvar

Class #2: Anchep

Number of samples : 77

Classification Accuracy : 97.4% (75/77)

Most confused with : Ancmit

Class #3: Anclyo

Number of samples : 8

Classification Accuracy : 87.5% (7/8)

Most confused with : Micund, Anchep, Lagrho

Class #4: Ancmit

Number of samples : 113

Classification Accuracy : 97.35% (110/113)

Most confused with : Anchep

Class #5: Ancomm

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Lutsyn, Lutcam

Class #6: Arcpro

Number of samples : 5

Classification Accuracy : 60% (3/5)

Most confused with : Stecap

Class #7: Arifel

Number of samples : 60

Classification Accuracy : 100% (60/60)

Most confused with : Micund

Class #8: Bagmar

Number of samples : 73

Classification Accuracy : 100% (73/73)

Most confused with : Larfas, Pogcro, Arifel, Scioce

Class #9: Baichr

Number of samples : 99

Classification Accuracy : 100% (99/99)

Most confused with : Stelan, Menber

Class #10: Baichr2

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Peppar, Lagrho, Leixan, Scioce, Lutcam, Anchep

Class #11: Batsop

Number of samples : 8

Classification Accuracy : 100% (8/8)

Most confused with : Evolyr, Citspi

Class #12: Brepat

Number of samples : 82

Classification Accuracy : 97.56% (80/82)

Most confused with : Harjag, Chlchr

Class #13: Carcry

Number of samples : 4

Classification Accuracy : 100% (4/4)

Most confused with : Chlchr, Lutgri

Class #14: Carhip

Number of samples : 1

Classification Accuracy : 0% (0/1)

Most confused with : Anchep, Mugspp, Chlchr, Eucspp

Class #15: Carhip2

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Baichr, Citspi, Ctebol

Class #16: Cenphi

Number of samples : 11

Classification Accuracy : 90.91% (10/11)

Most confused with : Censtr, Lutgri

Class #17: Censtr

Number of samples : 10

Classification Accuracy : 70% (7/10)

Most confused with : Lutgri

Class #18: Chafab

Number of samples : 8

Classification Accuracy : 87.5% (7/8)

Most confused with : Leixan, Chlchr, Lagrho

Class #19: Chasab

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Chlchr

Class #20: Chisch

Number of samples : 5

Classification Accuracy : 20% (1/5)

Most confused with : Ancmit, Eucspp

Class #21: Chlchr

Number of samples : 111

Classification Accuracy : 97.3% (108/111)

Most confused with : Lutgri

Class #22: Citmac

Number of samples : 56

Classification Accuracy : 78.57% (44/56)

Most confused with : Citspi

Class #23: Citspi

Number of samples : 105

Classification Accuracy : 79.05% (83/105)

Most confused with : Citmac

Class #24: Citspp

Number of samples : 5

Classification Accuracy : 40% (2/5)

Most confused with : Citmac

Class #25: Cryros

Number of samples : 3

Classification Accuracy : 100% (3/3)

Most confused with : Tracar, Chlchr, Nicust, Dorpet, Anchep

Class #26: Ctebol

Number of samples : 80

Classification Accuracy : 92.5% (74/80)

Most confused with : Sympla

Class #27: Cynare

Number of samples : 109

Classification Accuracy : 99.08% (108/109)

Most confused with : Cynneb

Class #28: Cynneb

Number of samples : 183

Classification Accuracy : 99.45% (182/183)

Most confused with : Mename

Class #29: Cypvar

Number of samples : 105

Classification Accuracy : 96.19% (101/105)

Most confused with : Poelat

Class #30: Dipfor

Number of samples : 12

Classification Accuracy : 100% (12/12)

Most confused with : Lutgri

Class #31: Dormac

Number of samples : 2

Classification Accuracy : 0% (0/2)

Most confused with : Poelat, Cypvar, Citspi, Fungra, Funsim, Ctebol, Evolyr, Goboce, Menber

Class #32: Dorpet

Number of samples : 27

Classification Accuracy : 85.19% (23/27)

Most confused with : Harjag

Class #33: Erosma

Number of samples : 8

Classification Accuracy : 87.5% (7/8)

Most confused with : Lagrho, Sympla

Class #34: Eucarg

Number of samples : 12

Classification Accuracy : 50% (6/12)

Most confused with : Eucspp

Class #35: Eucgul

Number of samples : 14

Classification Accuracy : 57.14% (8/14)

Most confused with : Eucspp

Class #36: Euchar

Number of samples : 9

Classification Accuracy : 33.33% (3/9)

Most confused with : Eucspp, Mugcur

Class #37: Eucmel

Number of samples : 34

Classification Accuracy : 91.18% (31/34)

Most confused with : Eucspp

Class #38: Eucspp

Number of samples : 113

Classification Accuracy : 79.65% (90/113)

Most confused with : Eucgul

Class #39: Evolyr

Number of samples : 17

Classification Accuracy : 82.35% (14/17)

Most confused with : Ctebol

Class #40: Fun

Number of samples : 7

Classification Accuracy : 100% (7/7)

Most confused with : Adixen, Cypvar, Ctebol, Sphpar

Class #41: Fungra

Number of samples : 121

Classification Accuracy : 93.39% (113/121)

Most confused with : Funsim

Class #42: Funmaj

Number of samples : 2

Classification Accuracy : 0% (0/2)

Most confused with : Funsim

Class #43: Funpul

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Funsim, Ctebol, Fungra

Class #44: Funsim

Number of samples : 96

Classification Accuracy : 95.83% (92/96)

Most confused with : Fungra

Class #45: Gamhol

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Mename, Citmac

Class #46: Gerreidae

Number of samples : 5

Classification Accuracy : 20% (1/5)

Most confused with : Eucmel

Class #47: Gobbos

Number of samples : 20

Classification Accuracy : 95% (19/20)

Most confused with : Citmac, Citspi, Baichr

Class #48: Goboce

Number of samples : 15

Classification Accuracy : 80% (12/15)

Most confused with : Ctebol

Class #49: Gobstr

Number of samples : 4

Classification Accuracy : 75% (3/4)

Most confused with : Lagrho, Scaridae, Sympla, Ctebol, Mugspp, Olisau

Class #50: Harjag

Number of samples : 121

Classification Accuracy : 96.69% (117/121)

Most confused with : Dorpet

Class #51: Hypmee

Number of samples : 5

Classification Accuracy : 80% (4/5)

Most confused with : Scioce, Leixan

Class #52: Lagrho

Number of samples : 137

Classification Accuracy : 87.59% (120/137)

Most confused with : Stecap

Class #53: Larfas

Number of samples : 46

Classification Accuracy : 91.3% (42/46)

Most confused with : Leixan

Class #54: Leixan

Number of samples : 128

Classification Accuracy : 99.22% (127/128)

Most confused with : Scioce, Micund, Mename

Class #55: Lucpar

Number of samples : 5

Classification Accuracy : 80% (4/5)

Most confused with : Cypvar, Fungra, Citspi, Ctebol

Class #56: Lutcam

Number of samples : 72

Classification Accuracy : 97.22% (70/72)

Most confused with : Lutsyn

Class #57: Lutgri

Number of samples : 147

Classification Accuracy : 95.24% (140/147)

Most confused with : Censtr

Class #58: Lutsyn

Number of samples : 130

Classification Accuracy : 96.92% (126/130)

Most confused with : Lagrho

Class #59: Memmar

Number of samples : 81

Classification Accuracy : 91.36% (74/81)

Most confused with : Menber

Class #60: Mename

Number of samples : 151

Classification Accuracy : 100% (151/151)

Most confused with : Cynneb

Class #61: Menber

Number of samples : 149

Classification Accuracy : 91.95% (137/149)

Most confused with : Memmar

Class #62: Menlit

Number of samples : 6

Classification Accuracy : 66.67% (4/6)

Most confused with : Mename

Class #63: Menpen

Number of samples : 44

Classification Accuracy : 86.36% (38/44)

Most confused with : Lagrho

Class #64: Menspp

Number of samples : 4

Classification Accuracy : 100% (4/4)

Most confused with : Mename

Class #65: Micund

Number of samples : 177

Classification Accuracy : 98.31% (174/177)

Most confused with : Leixan

Class #66: Mugcep

Number of samples : 87

Classification Accuracy : 98.85% (86/87)

Most confused with : Mugcur

Class #67: Mugcur

Number of samples : 77

Classification Accuracy : 97.4% (75/77)

Most confused with : Mugcep

Class #68: Mugspp

Number of samples : 38

Classification Accuracy : 97.37% (37/38)

Most confused with : Mugcep, Lagrho

Class #69: Mycmic

Number of samples : 10

Classification Accuracy : 100% (10/10)

Most confused with : Lutgri

Class #70: Myrpun

Number of samples : 7

Classification Accuracy : 100% (7/7)

Most confused with : Lagrho, Sympla

Class #71: Nicust

Number of samples : 13

Classification Accuracy : 84.62% (11/13)

Most confused with : Chlchr

Class #72: Olisau

Number of samples : 6

Classification Accuracy : 16.67% (1/6)

Most confused with : Mugspp

Class #73: Opiogl

Number of samples : 4

Classification Accuracy : 25% (1/4)

Most confused with : Dorpet

Class #74: Opsbet

Number of samples : 8

Classification Accuracy : 100% (8/8)

Most confused with : Mugspp, Mugcep

Class #75: Ortchr

Number of samples : 2

Classification Accuracy : 0% (0/2)

Most confused with : Lutsyn, Lutcam

Class #76: Paralb

Number of samples : 36

Classification Accuracy : 86.11% (31/36)

Most confused with : Lagrho

Class #77: Parlet

Number of samples : 7

Classification Accuracy : 42.86% (3/7)

Most confused with : Paralb

Class #78: Pepbur

Number of samples : 12

Classification Accuracy : 83.33% (10/12)

Most confused with : Chlchr, Selvom, Eucspp, Synfoe

Class #79: Peppar

Number of samples : 36

Classification Accuracy : 97.22% (35/36)

Most confused with : Eucspp, Lutcam

Class #80: Poelat

Number of samples : 39

Classification Accuracy : 97.44% (38/39)

Most confused with : Fungra, Cypvar

Class #81: Pogcro

Number of samples : 4

Classification Accuracy : 100% (4/4)

Most confused with : Bagmar, Scioce

Class #82: Pomsal

Number of samples : 4

Classification Accuracy : 50% (2/4)

Most confused with : Chlchr, Mugcur

Class #83: Priros

Number of samples : 7

Classification Accuracy : 100% (7/7)

Most confused with : Lutcam

Class #84: Pritri

Number of samples : 11

Classification Accuracy : 45.45% (5/11)

Most confused with : Eucspp, Mename

Class #85: Raccan

Number of samples : 2

Classification Accuracy : 50% (1/2)

Most confused with : Chlchr, Lutgri, Lutsyn, Mugcur

Class #86: Scaridae

Number of samples : 5

Classification Accuracy : 80% (4/5)

Most confused with : Lagrho

Class #87: Scioce

Number of samples : 110

Classification Accuracy : 98.18% (108/110)

Most confused with : Lutcam, Cynare

Class #88: Selvom

Number of samples : 36

Classification Accuracy : 86.11% (31/36)

Most confused with : Menber, Eucspp

Class #89: Sparad

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Mugcur, Menber, Eucgul, Menpen, Eucspp, Tracar, Trilep

Class #90: Sparub

Number of samples : 4

Classification Accuracy : 50% (2/4)

Most confused with : Lutgri

Class #91: Sphbar

Number of samples : 2

Classification Accuracy : 50% (1/2)

Most confused with : Synfoe

Class #92: Sphgua

Number of samples : 10

Classification Accuracy : 100% (10/10)

Most confused with : Chlchr, Cenphi

Class #93: Sphpar

Number of samples : 57

Classification Accuracy : 96.49% (55/57)

Most confused with : Ctebol

Class #94: Stecap

Number of samples : 114

Classification Accuracy : 98.25% (112/114)

Most confused with : Lutsyn, Lagrho, Memmar

Class #95: Stelan

Number of samples : 22

Classification Accuracy : 100% (22/22)

Most confused with : Erosma, Menber

Class #96: Stelan2

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Peppar, Leixan, Priros

Class #97: Strtim

Number of samples : 2

Classification Accuracy : 100% (2/2)

Most confused with : Lutsyn, Lagrho, Lutgri, Censtr, Cynneb, Paralb, Scioce

Class #98: Sygspp

Number of samples : 1

Classification Accuracy : 0% (0/1)

Most confused with : Cynare, Cynneb, Ancmit, Dipfor, Gamhol, Menber, Mugcep

Class #99: Sympla

Number of samples : 65

Classification Accuracy : 96.92% (63/65)

Most confused with : Ctebol

Class #100: Synfoe

Number of samples : 73

Classification Accuracy : 97.26% (71/73)

Most confused with : Lutgri, Mugcep

Class #101: Synlou

Number of samples : 1

Classification Accuracy : 0% (0/1)

Most confused with : Chlchr, Anchep, Arcpro, Lagrho, Memmar, Menber, Synfoe

Class #102: Tracar

Number of samples : 16

Classification Accuracy : 75% (12/16)

Most confused with : Chlchr

Class #103: Trilep

Number of samples : 4

Classification Accuracy : 100% (4/4)

Most confused with : Tracar, Synfoe

Class #104: Trimac

Number of samples : 5

Classification Accuracy : 40% (2/5)

Most confused with : Menber

Intra-Class Statistics

This section presents embedding statistics on all possible image pairs (excluding identities) for each class.
(Beware: This does not include ICP steps).

Otolith Image Processing Library on Dauphin Island Sea Lab

OIPL Statistics

Otolith Classification With Neural Embeddings & ICP

Table of Contents

Individual Statistics

Class #0: Achlin

Class #1: Adixen

Class #2: Anchep

Class #3: Anclyo

Class #4: Ancmit

Class #5: Ancomm

Class #6: Arcpro

Class #7: Arifel

Class #8: Bagmar

Class #9: Baichr

Class #10: Baichr2

Class #11: Batsop

Class #12: Brepat

Class #13: Carcry

Class #14: Carhip

Class #15: Carhip2

Class #16: Cenphi

Class #17: Censtr

Class #18: Chafab

Class #19: Chasab

Class #20: Chisch

Class #21: Chlchr

Class #22: Citmac

Class #23: Citspi

Class #24: Citspp

Class #25: Cryros

Class #26: Ctebol

Class #27: Cynare

Class #28: Cynneb

Class #29: Cypvar

Class #30: Dipfor

Class #31: Dormac

Class #32: Dorpet

Class #33: Erosma

Class #34: Eucarg

Class #35: Eucgul

Class #36: Euchar

Class #37: Eucmel

Class #38: Eucspp

Class #39: Evolyr

Class #40: Fun

Class #41: Fungra

Class #42: Funmaj

Class #43: Funpul

Class #44: Funsim

Class #45: Gamhol

Class #46: Gerreidae

Class #47: Gobbos

Class #48: Goboce

Class #49: Gobstr

Class #50: Harjag

Class #51: Hypmee

Class #52: Lagrho

Class #53: Larfas

Class #54: Leixan

Class #55: Lucpar

Class #56: Lutcam

Class #57: Lutgri

Class #58: Lutsyn

Class #59: Memmar

Class #60: Mename

Class #61: Menber

Class #62: Menlit

Class #63: Menpen

Class #64: Menspp

Class #65: Micund

Class #66: Mugcep

Class #67: Mugcur

Class #68: Mugspp

Class #69: Mycmic

Class #70: Myrpun

Class #71: Nicust

Class #72: Olisau

Class #73: Opiogl

Class #74: Opsbet