Fish Ovary
Each folder starting by oocytes_
contain a dataset. See the README.md
file inside each folder for further information about each one.
The data are in the files XX_R.dat
(e.g. oocytes_merluccius_states_2f_R.dat
). These data are texture features of oocytes extracted from histological images of fish ovary of species Merluccius and Trisopterus. There are 4 data sets (each one corresponds to a different combination of texture features, as described in the reference below):
Set | Instances | Inputs | Classes |
---|---|---|---|
Merluccius States | 1022 | 25 | 3 |
Merluccius Nucleus | 1022 | 41 | 2 |
Trisopterus States | 912 | 25 | 3 |
Trisopterus Nucleus | 912 | 25 | 2 |
The values of each input are standarized with zero mean and standard deviation one. The classes in the Merluccius and Trisopterus state data sets are ordinal, i.e., there is an order from class 0 to class 1 and from class 1 to class 2, which the three classes are states of development of oocytes. Therefore, these two data sets are ordinal classification problems.
The folders starting by reinhardtius_
contain data from species Reinhardtius hippoglossoides with six developmental states: primary growth (pg), alveolli cortical (ac), vit1 (vitellogenetic sub-level #1), vit2 (vitellogenetic sub-level #2), vit3 and vit4: the file reinhardtius_hippoglossoides-leave-one-image-out.tar.gz has 16 directories, each with the training, validation and test sets of each one of 16 images of this species using a Leave-One-Image-Out validation, i.e., for the i-th trial, the training and validation sets contain patterns from all the images but the i-th image, and the test set contains patterns from the i-th image.
The file reinhardtius_hippoglossoides-mixed-images.tar.gz contains 10 directories, each one with a training, validation and test files: both three files contain patterns from the 16 images, in order to develop a 10-fold cross validation. All the patterns have 25 inputs: the 10 first are grey level texture features (Local Binary Patterns); the last 15 are chromatic features. The following table specifies the instances and number of patterns of each class for each image
Image | Instances | pg | ac | vit1 | vit2 | vit3 | vit4 |
---|---|---|---|---|---|---|---|
1 | 668 | 239 | 124 | 177 | 128 | 0 | 0 |
2 | 685 | 539 | 146 | 0 | 0 | 0 | 0 |
3 | 820 | 470 | 184 | 166 | 0 | 0 | 0 |
4 | 507 | 195 | 87 | 86 | 40 | 99 | 0 |
5 | 840 | 330 | 167 | 158 | 185 | 0 | 0 |
6 | 97 | 27 | 16 | 19 | 0 | 0 | 35 |
7 | 472 | 0 | 103 | 158 | 211 | 0 | 0 |
8 | 919 | 443 | 128 | 206 | 140 | 2 | 0 |
9 | 544 | 258 | 13 | 103 | 42 | 128 | 0 |
10 | 539 | 98 | 68 | 68 | 1 | 304 | 0 |
11 | 518 | 0 | 204 | 314 | 0 | 0 | 0 |
12 | 408 | 66 | 76 | 71 | 173 | 22 | 0 |
13 | 205 | 0 | 54 | 45 | 0 | 70 | 36 |
14 | 153 | 65 | 18 | 22 | 0 | 0 | 48 |
15 | 191 | 91 | 35 | 22 | 0 | 0 | 43 |
16 | 349 | 150 | 64 | 35 | 0 | 7 | 93 |
Sum | 2971 | 1487 | 1650 | 920 | 632 | 255 |
Reference
@article{rufino13,
author="{E. Gonz\'alez-Rufino and P. Carri\'on and E. Cernadas and M. Fern\'andez-Delgado and R. Dom\'inguez-Petit}",
journal={Pattern Recognition},
pages = {2391—2407},
title = "{Exhaustive comparison of colour texture features and classification methods to discriminate cells categories in histological images of fish ovary}",
volume = {46},
year = {2013},
}
@article{perez15,
author="{M. P\'erez-Ortiz and M. Fern\'andez-Delgado and E. Cernadas and R. Dom\'inguez-Petit and P.A. Guti\'errez and C. Herv\'as-Mart\'inez}",
journal={Neural Processing Letters},
title = "{On the use of nominal and ordinal classifiers for the discrimination of states of development in fish oocytes}",
pages = {1--16},
url={http://dx.doi.org/10.1007/s11063-015-9476-8},
doi = {10.1007/s11063-015-9476-8},
issn="1573-773X",
}
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
You can use this dataset on your publication as long as you include a citation to the reference on this page. When including a link to this dataset, please use this page instead of linking the file directly.