Update Readme to move download instructions at the top
Browse files
README.md
CHANGED
@@ -49,54 +49,35 @@ Image with caption:
|
|
49 |
-->
|
50 |
|
51 |
# Dataset Card for Fish-Visual Trait Analysis (Fish-Vista)
|
52 |
-
|
53 |
-
## Dataset Details
|
54 |
-
|
55 |
-
### Dataset Description
|
56 |
-
|
57 |
-
<!--
|
58 |
-
- **Curated by:** list curators (authors for _data_ citation, moved up)
|
59 |
-
- **Language(s) (NLP):** [More Information Needed]
|
60 |
-
<!-- Provide the basic links for the dataset. These will show up on the sidebar to the right of your dataset card ("Curated by" too). -->
|
61 |
-
<!--
|
62 |
-
- **Homepage:**
|
63 |
-
- **Repository:** [related project repo]
|
64 |
-
- **Paper:**
|
65 |
-
-->
|
66 |
-
|
67 |
-
<!-- Provide a longer summary of what this dataset is. -->
|
68 |
-
|
69 |
-
The Fish-Visual Trait Analysis (Fish-Vista) dataset is a large, annotated collection of 60K fish images spanning 1900 different species; it supports several challenging and biologically relevant tasks including species classification, trait identification, and trait segmentation. These images have been curated through a sophisticated data processing pipeline applied to a cumulative set of images obtained from various museum collections. Fish-Vista provides fine-grained labels of various visual traits present in each image. It also offers pixel-level annotations of 9 different traits for 2427 fish images, facilitating additional trait segmentation and localization tasks.
|
70 |
-
|
71 |
-
The Fish Vista dataset consists of museum fish images from [Great Lakes Invasives Network (GLIN)](https://greatlakesinvasives.org/portal/index.php), [iDigBio](https://www.idigbio.org/), and [Morphbank](https://www.morphbank.net/) databases. We acquired these images, along with associated metadata including the scientific species names, the taxonomical family the species belong to, and licensing information, from the [Fish-AIR repository](https://fishair.org/).
|
72 |
-
|
73 |
|![Figure 1](https://huggingface.co/datasets/imageomics/fish-vista/resolve/main/metadata/figures/FishVista.png)|
|
74 |
|:--|
|
75 |
|**Figure 1.** A schematic representation of the different tasks in Fish-Vista Dataset. |
|
76 |
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
|
89 |
-
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
--->
|
95 |
-
|
96 |
|
97 |
-
|
98 |
|
99 |
-
|
|
|
|
|
|
|
|
|
100 |
|
101 |
## Dataset Structure
|
102 |
|
@@ -142,31 +123,6 @@ English
|
|
142 |
data-bib.bib
|
143 |
```
|
144 |
|
145 |
-
### Instructions for downloading dataset
|
146 |
-
<!-- [Add instructions for downloading images here]
|
147 |
-
-->
|
148 |
-
* Install [Git LFS](https://docs.github.com/en/repositories/working-with-files/managing-large-files/installing-git-large-file-storage)
|
149 |
-
* Git clone the fish-vista repository
|
150 |
-
* Run the following commands in a **terminal**:
|
151 |
-
```bash
|
152 |
-
git clone https://huggingface.co/datasets/imageomics/fish-vista
|
153 |
-
cd fish-vista
|
154 |
-
```
|
155 |
-
* Run the following commands to move all chunked images to a single directory:
|
156 |
-
```bash
|
157 |
-
mkdir AllImages
|
158 |
-
find Images -type f -exec mv -v {} AllImages \;
|
159 |
-
rm -rf Images
|
160 |
-
mv AllImages Images
|
161 |
-
```
|
162 |
-
|
163 |
-
* You should now have all the images in the *Images* directory
|
164 |
-
|
165 |
-
* Run the following commands to download and process copyrighted images
|
166 |
-
```bash
|
167 |
-
python download_and_process_nd_images.py --save_dir Images
|
168 |
-
```
|
169 |
-
* This will download and process the CC-BY-ND images that we do not provide in the *Images* folder
|
170 |
|
171 |
### Data Instances
|
172 |
|
@@ -218,13 +174,57 @@ CSV Columns are as follows:
|
|
218 |
- `barbel`: Presence/absence of the barbel trait. NA for the classification (FV-419) dataset, since it is only used for identification. 1 indicates presence and 0 indicates absence. This is used for trait identification.
|
219 |
- `multiple_dorsal_fin`: Presence/absence of the dorsal fin trait. NA for the classification (FV-419) dataset, since it is only used for identification. 1 indicates presence, 0 indicates absence and -1 indicates unknown. This is used for trait identification.
|
220 |
|
221 |
-
**Note:**
|
222 |
|
223 |
|
224 |
### Data Splits
|
225 |
|
226 |
For each task (or subset), the split is indicated by the CSV name (e.g., `classification_<split>.csv`). More information is provided in [Data Instances](#data-instances), above.
|
227 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
228 |
## Dataset Creation
|
229 |
|
230 |
### Curation Rationale
|
|
|
49 |
-->
|
50 |
|
51 |
# Dataset Card for Fish-Visual Trait Analysis (Fish-Vista)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
52 |
|![Figure 1](https://huggingface.co/datasets/imageomics/fish-vista/resolve/main/metadata/figures/FishVista.png)|
|
53 |
|:--|
|
54 |
|**Figure 1.** A schematic representation of the different tasks in Fish-Vista Dataset. |
|
55 |
|
56 |
+
## Instructions for downloading dataset and images
|
57 |
+
<!-- [Add instructions for downloading images here]
|
58 |
+
-->
|
59 |
+
* Install [Git LFS](https://docs.github.com/en/repositories/working-with-files/managing-large-files/installing-git-large-file-storage)
|
60 |
+
* Git clone the fish-vista repository
|
61 |
+
* Run the following commands in a **terminal**:
|
62 |
+
```bash
|
63 |
+
git clone https://huggingface.co/datasets/imageomics/fish-vista
|
64 |
+
cd fish-vista
|
65 |
+
```
|
66 |
+
* Run the following commands to move all chunked images to a single directory:
|
67 |
+
```bash
|
68 |
+
mkdir AllImages
|
69 |
+
find Images -type f -exec mv -v {} AllImages \;
|
70 |
+
rm -rf Images
|
71 |
+
mv AllImages Images
|
72 |
+
```
|
|
|
|
|
73 |
|
74 |
+
* You should now have all the images in the *Images* directory
|
75 |
|
76 |
+
* Run the following commands to download and process copyrighted images
|
77 |
+
```bash
|
78 |
+
python download_and_process_nd_images.py --save_dir Images
|
79 |
+
```
|
80 |
+
* This will download and process the CC-BY-ND images that we do not provide in the *Images* folder
|
81 |
|
82 |
## Dataset Structure
|
83 |
|
|
|
123 |
data-bib.bib
|
124 |
```
|
125 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
126 |
|
127 |
### Data Instances
|
128 |
|
|
|
174 |
- `barbel`: Presence/absence of the barbel trait. NA for the classification (FV-419) dataset, since it is only used for identification. 1 indicates presence and 0 indicates absence. This is used for trait identification.
|
175 |
- `multiple_dorsal_fin`: Presence/absence of the dorsal fin trait. NA for the classification (FV-419) dataset, since it is only used for identification. 1 indicates presence, 0 indicates absence and -1 indicates unknown. This is used for trait identification.
|
176 |
|
|
|
177 |
|
178 |
|
179 |
### Data Splits
|
180 |
|
181 |
For each task (or subset), the split is indicated by the CSV name (e.g., `classification_<split>.csv`). More information is provided in [Data Instances](#data-instances), above.
|
182 |
+
|
183 |
+
|
184 |
+
## Dataset Details
|
185 |
+
|
186 |
+
### Dataset Description
|
187 |
+
|
188 |
+
<!--
|
189 |
+
- **Curated by:** list curators (authors for _data_ citation, moved up)
|
190 |
+
- **Language(s) (NLP):** [More Information Needed]
|
191 |
+
<!-- Provide the basic links for the dataset. These will show up on the sidebar to the right of your dataset card ("Curated by" too). -->
|
192 |
+
<!--
|
193 |
+
- **Homepage:**
|
194 |
+
- **Repository:** [related project repo]
|
195 |
+
- **Paper:**
|
196 |
+
-->
|
197 |
+
|
198 |
+
<!-- Provide a longer summary of what this dataset is. -->
|
199 |
+
|
200 |
+
The Fish-Visual Trait Analysis (Fish-Vista) dataset is a large, annotated collection of 60K fish images spanning 1900 different species; it supports several challenging and biologically relevant tasks including species classification, trait identification, and trait segmentation. These images have been curated through a sophisticated data processing pipeline applied to a cumulative set of images obtained from various museum collections. Fish-Vista provides fine-grained labels of various visual traits present in each image. It also offers pixel-level annotations of 9 different traits for 2427 fish images, facilitating additional trait segmentation and localization tasks.
|
201 |
+
|
202 |
+
The Fish Vista dataset consists of museum fish images from [Great Lakes Invasives Network (GLIN)](https://greatlakesinvasives.org/portal/index.php), [iDigBio](https://www.idigbio.org/), and [Morphbank](https://www.morphbank.net/) databases. We acquired these images, along with associated metadata including the scientific species names, the taxonomical family the species belong to, and licensing information, from the [Fish-AIR repository](https://fishair.org/).
|
203 |
+
|
204 |
+
|
205 |
+
<!--This dataset card has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/datasetcard_template.md?plain=1), and further altered to suit Imageomics Institute needs.-->
|
206 |
+
|
207 |
+
|
208 |
+
### Supported Tasks and Leaderboards
|
209 |
+
<!--[Add some more description. could replace graphs with tables]-->
|
210 |
+
|
211 |
+
|![Figure 2](https://huggingface.co/datasets/imageomics/fish-vista/resolve/main/metadata/figures/clf_imbalance.png)|
|
212 |
+
|:--|
|
213 |
+
|**Figure 2.** Comparison of the fine-grained classification performance of different imbalanced classification methods. |
|
214 |
+
|
215 |
+
|![Figure 3](https://huggingface.co/datasets/imageomics/fish-vista/resolve/main/metadata/figures/IdentificationOriginalResults.png)|
|
216 |
+
|:--|
|
217 |
+
|**Figure 3.** Trait identification performance of different multi-label classification methods. |
|
218 |
+
|
219 |
+
|
220 |
+
<!---
|
221 |
+
This dataset card aims to be a base template for new datasets. It has been generated using [this raw template](https://github.com/huggingface/huggingface_hub/blob/main/src/huggingface_hub/templates/datasetcard_template.md?plain=1).
|
222 |
+
--->
|
223 |
+
|
224 |
+
### Languages
|
225 |
+
|
226 |
+
English
|
227 |
+
|
228 |
## Dataset Creation
|
229 |
|
230 |
### Curation Rationale
|