Datasets:
Rename deduplication notebook to reflect its order of operation
Browse files
notebooks/{dedupe_full_jiggins_DL.ipynb → Data-gen-1-2.ipynb}
RENAMED
@@ -15,7 +15,7 @@
|
|
15 |
"source": [
|
16 |
"## Summary\n",
|
17 |
"\n",
|
18 |
-
"We explored our downloaded images in `
|
19 |
"1. The \"patch\" record (record 3477891) which required alignment in `record_3477891_media_occurrence.ipynb`, resulting in `Jiggins_Zenodo_Img_Master_3477891Patch.csv` used for download, is a duplication of all listed records. Additionally, it introduced extra copies of 4 images with incorrect labels for their views (i.e., dorsal vs ventral).\n",
|
20 |
" - Solution: This record will be removed.\n",
|
21 |
"2. The remaining duplication (430 images x2) comes from 5 records: [4291095](https://zenodo.org/records/4291095), [2813153](https://zenodo.org/records/2813153), [5526257](https://zenodo.org/records/5526257), [2553977](https://zenodo.org/records/2553977), and [2552371](https://zenodo.org/records/2552371).\n",
|
@@ -28,9 +28,10 @@
|
|
28 |
"3. Looking at images that were not duplicates, there are clearly still multiple images of the same specimen from the same perspective (eg., two dorsal images) that are also of the same file type (eg., both jpgs). Duplication was checked at the pixel level, so there is no guarantee that they are truly different images, but due to the scope of this collection (over 20 years), it does not seem unlikely that multiple images could have been taken of the same specimen. \n",
|
29 |
" - Suggestion/Solution: When determining splits using this data, follow these steps:\n",
|
30 |
" 1. Ensure you are looking at only **_one_** view (eg., dorsal or ventral).\n",
|
31 |
-
" 2.
|
32 |
-
" 3.
|
33 |
-
" 4.
|
|
|
34 |
"\n",
|
35 |
"\n",
|
36 |
"## Generate \"Clean\" Master Metadata File\n",
|
|
|
15 |
"source": [
|
16 |
"## Summary\n",
|
17 |
"\n",
|
18 |
+
"We explored our downloaded images in `EDA-DL-0-3.ipynb`, and established the following causes of duplication along with remedies:\n",
|
19 |
"1. The \"patch\" record (record 3477891) which required alignment in `record_3477891_media_occurrence.ipynb`, resulting in `Jiggins_Zenodo_Img_Master_3477891Patch.csv` used for download, is a duplication of all listed records. Additionally, it introduced extra copies of 4 images with incorrect labels for their views (i.e., dorsal vs ventral).\n",
|
20 |
" - Solution: This record will be removed.\n",
|
21 |
"2. The remaining duplication (430 images x2) comes from 5 records: [4291095](https://zenodo.org/records/4291095), [2813153](https://zenodo.org/records/2813153), [5526257](https://zenodo.org/records/5526257), [2553977](https://zenodo.org/records/2553977), and [2552371](https://zenodo.org/records/2552371).\n",
|
|
|
28 |
"3. Looking at images that were not duplicates, there are clearly still multiple images of the same specimen from the same perspective (eg., two dorsal images) that are also of the same file type (eg., both jpgs). Duplication was checked at the pixel level, so there is no guarantee that they are truly different images, but due to the scope of this collection (over 20 years), it does not seem unlikely that multiple images could have been taken of the same specimen. \n",
|
29 |
" - Suggestion/Solution: When determining splits using this data, follow these steps:\n",
|
30 |
" 1. Ensure you are looking at only **_one_** view (eg., dorsal or ventral).\n",
|
31 |
+
" 2. Ensure you are only looking one file type (`raw`, `jpg`, or `tif`).\n",
|
32 |
+
" 3. Reduce to only unique `CAMID`s.\n",
|
33 |
+
" 4. Generate splits as desired.\n",
|
34 |
+
" 5. Add images to splits based on matched `CAMID`s if multiple views are desired (i.e., if a `CAMID` is present in the dorsal training, add the matching ventral image to the training set).\n",
|
35 |
"\n",
|
36 |
"\n",
|
37 |
"## Generate \"Clean\" Master Metadata File\n",
|