ADMorph:a 3D Digital Microfossil Morphology Dataset for Deep Learning

A random sample of shapes (one per class) was chosen from the experimental dataset. (a) Gualepis. (b) Youngolepis. (c) Psarolepis. (d) Guiyu. (e) Galeaspida indet. (f) Acanthodii indet. (g) Placodermi indet. (h) Nostolepis. (i) Parathelodus wangi. (j) P. liaokuoensis. (k) P. asiaticus. Scale bar, a-c, e-k = 0.2 mm, d = 1 mm.

About the ADMorph Dataset

ADMorph dataset, a large-scale 3D microfossil model dataset, contains over ten thousand 3D virtual models of early fish bones, teeth and scales from 5 main classes of early fishes (agnathans, placoderms, osteichthyans, chondrichthyans, and acanthodians). So far, we selected 2,010 labeled scale virtual models among the dataset to open. We labeled the early fish scales into different categories according to their morphology and histology. The virtual models contained agnathans (Galeaspida indet, 105; Parathelodus asiaticus, 57; P. liaokuoensis, 100; P. wangi, 100), placoderms (Placodermi indet 106), osteichthyans (Guiyu,145; Youngolepis, 163; Psarolepis, 170), chondrichthyans (Gualepis, 888), and acanthodians (Acanthodii indet, 34; Nostolepis, 142).

Download ADMorph Dataset

ADMorph Dataset.zip:
this ZIP file contains microfossil models from the 11 categories used to train the deep network. The microfossil models are completely labeled by paleontologists.

Citation

If you find this dataset useful, please cite the following paper:
Y.Hou, X.Cui, M.Ku, S.Jin, R.HASIMOTO-BELTRAN, Q.Hua and M.Zhu
ADMorph: a 3D Digital Microfossil Morphology Dataset for Deep Learning
IEEE ACCESS

Acknowledgements

This research was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA19050102, XDB26000000), the National Natural Science Foundation of China (41530102), the Key Research Program of Frontier Sciences, CAS (QYZDJ-SSW-DQC002).