Upload folder using huggingface_hub
Browse files- README.md +118 -0
- data/contrast.zip +3 -0
- data/gaussian_noise.zip +3 -0
- data/impulse_noise.zip +3 -0
- data/jpeg_compression.zip +3 -0
- data/motion_blur.zip +3 -0
- data/pixelate.zip +3 -0
- data/spatter.zip +3 -0
- data/test.zip +3 -0
- data/train.zip +3 -0
- resisc45.py +108 -0
README.md
ADDED
@@ -0,0 +1,118 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
dataset_info:
|
3 |
+
features:
|
4 |
+
- name: image
|
5 |
+
dtype: image
|
6 |
+
- name: label
|
7 |
+
dtype:
|
8 |
+
class_label:
|
9 |
+
names:
|
10 |
+
'0': airplane
|
11 |
+
'1': airport
|
12 |
+
'2': baseball diamond
|
13 |
+
'3': basketball court
|
14 |
+
'4': beach
|
15 |
+
'5': bridge
|
16 |
+
'6': chaparral
|
17 |
+
'7': church
|
18 |
+
'8': circular farmland
|
19 |
+
'9': cloud
|
20 |
+
'10': commercial area
|
21 |
+
'11': dense residential
|
22 |
+
'12': desert
|
23 |
+
'13': forest
|
24 |
+
'14': freeway
|
25 |
+
'15': golf course
|
26 |
+
'16': ground track field
|
27 |
+
'17': harbor
|
28 |
+
'18': industrial area
|
29 |
+
'19': intersection
|
30 |
+
'20': island
|
31 |
+
'21': lake
|
32 |
+
'22': meadow
|
33 |
+
'23': medium residential
|
34 |
+
'24': mobile home park
|
35 |
+
'25': mountain
|
36 |
+
'26': overpass
|
37 |
+
'27': palace
|
38 |
+
'28': parking lot
|
39 |
+
'29': railway
|
40 |
+
'30': railway station
|
41 |
+
'31': rectangular farmland
|
42 |
+
'32': river
|
43 |
+
'33': roundabout
|
44 |
+
'34': runway
|
45 |
+
'35': sea ice
|
46 |
+
'36': ship
|
47 |
+
'37': snowberg
|
48 |
+
'38': sparse residential
|
49 |
+
'39': stadium
|
50 |
+
'40': storage tank
|
51 |
+
'41': tennis court
|
52 |
+
'42': terrace
|
53 |
+
'43': thermal power station
|
54 |
+
'44': wetland
|
55 |
+
splits:
|
56 |
+
- name: train
|
57 |
+
num_bytes: 2383019
|
58 |
+
num_examples: 18900
|
59 |
+
- name: test
|
60 |
+
num_bytes: 781557
|
61 |
+
num_examples: 6300
|
62 |
+
- name: contrast
|
63 |
+
num_bytes: 831957
|
64 |
+
num_examples: 6300
|
65 |
+
- name: gaussian_noise
|
66 |
+
num_bytes: 907557
|
67 |
+
num_examples: 6300
|
68 |
+
- name: impulse_noise
|
69 |
+
num_bytes: 894957
|
70 |
+
num_examples: 6300
|
71 |
+
- name: jpeg_compression
|
72 |
+
num_bytes: 932757
|
73 |
+
num_examples: 6300
|
74 |
+
- name: motion_blur
|
75 |
+
num_bytes: 869757
|
76 |
+
num_examples: 6300
|
77 |
+
- name: pixelate
|
78 |
+
num_bytes: 831957
|
79 |
+
num_examples: 6300
|
80 |
+
- name: spatter
|
81 |
+
num_bytes: 819357
|
82 |
+
num_examples: 6300
|
83 |
+
download_size: 916981012
|
84 |
+
dataset_size: 9252875
|
85 |
+
---
|
86 |
+
|
87 |
+
# [RESISC45](https://www.tensorflow.org/datasets/catalog/resisc45)
|
88 |
+
|
89 |
+
## Overview
|
90 |
+
|
91 |
+
The **RESISC45 Classification Model** is a deep learning-based image classification model designed to identify and classify different types of scenes from satellite imagery. This dataset contains 45 classes of scenes, with each class representing a different type of land cover or land use.
|
92 |
+
|
93 |
+
## Usage
|
94 |
+
|
95 |
+
```python
|
96 |
+
from datasets import load_dataset
|
97 |
+
|
98 |
+
# Load the dataset
|
99 |
+
dataset = load_dataset('tanganke/resisc45')
|
100 |
+
```
|
101 |
+
|
102 |
+
## Dataset Information
|
103 |
+
|
104 |
+
The dataset is divided into the following splits:
|
105 |
+
|
106 |
+
- **Training set**: Contains 18,900 examples, used for model training.
|
107 |
+
- **Test set**: Contains 6,300 examples, used for model evaluation and benchmarking.
|
108 |
+
|
109 |
+
The dataset also includes the following augmented sets, which can be used for testing the model's robustness to various types of image corruptions:
|
110 |
+
|
111 |
+
- **Contrast-enhanced set**: Contains 6,300 examples with enhanced contrast for improved feature visibility.
|
112 |
+
- **Gaussian noise set**: Contains 6,300 examples where images have been corrupted with Gaussian noise.
|
113 |
+
- **Impulse noise set**: Contains 6,300 examples with impulse noise.
|
114 |
+
- **JPEG compression set**: Contains 6,300 examples where images have been compressed using JPEG encoding.
|
115 |
+
- **Motion blur set**: Contains 6,300 examples with motion blur applied.
|
116 |
+
- **Pixelate set**: Contains 6,300 examples where images have been pixelated.
|
117 |
+
- **Spatter set**: Contains 6,300 examples with spatter noise.
|
118 |
+
|
data/contrast.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:95b995aa5742a8bf5c99caec21e425ed62e8be5e7e2499b770e177a06442ceb0
|
3 |
+
size 65821198
|
data/gaussian_noise.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e3fbcc81eb278e18e9e8bc75a220d07caccc1c8ac8c55eed2d84c7f8aca804df
|
3 |
+
size 115764717
|
data/impulse_noise.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7422b2cde62e6697a3d0ef3f0727aca21b788ab7b45a1622a2b57f4c65c32332
|
3 |
+
size 124570638
|
data/jpeg_compression.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5e3c4d35e8b22f6db7e50f8394f7cd1e40dbbeec75f4cd92b27b9610b2d585fc
|
3 |
+
size 82819474
|
data/motion_blur.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8e372cae5fdc68296741d71742333cc39c05860faa463d37a734d52e14bd0e1d
|
3 |
+
size 72616247
|
data/pixelate.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:10ec683407ab63d57b7109fa9c78edf7a78b1e83572bfeddd052decee8e4e0c4
|
3 |
+
size 5677005
|
data/spatter.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:99bb1700ff22675d5521d919b59ad26dd6716c3e4d98ecb89fdb2cdef23f827f
|
3 |
+
size 108574042
|
data/test.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5955d0c47b55a178b30a3b6f5f8dff1447a3d2d6e018a6867ddb3d0e640de32b
|
3 |
+
size 85393782
|
data/train.zip
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0427735bc7efb681e9b680ee49839f53a1e3087d9734f6b507a057ab23be5f10
|
3 |
+
size 255743909
|
resisc45.py
ADDED
@@ -0,0 +1,108 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import datasets
|
2 |
+
from datasets.data_files import DataFilesDict
|
3 |
+
from datasets.packaged_modules.imagefolder.imagefolder import ImageFolder, ImageFolderConfig
|
4 |
+
|
5 |
+
logger = datasets.logging.get_logger(__name__)
|
6 |
+
|
7 |
+
|
8 |
+
class RESISC45(ImageFolder):
|
9 |
+
R"""
|
10 |
+
RESISC45 dataset for image classification.
|
11 |
+
"""
|
12 |
+
|
13 |
+
BUILDER_CONFIG_CLASS = ImageFolderConfig
|
14 |
+
BUILDER_CONFIGS = [
|
15 |
+
ImageFolderConfig(
|
16 |
+
name="default",
|
17 |
+
features=("images", "labels"),
|
18 |
+
data_files=DataFilesDict(
|
19 |
+
{
|
20 |
+
split: f"data/{split}.zip"
|
21 |
+
for split in ["train", "test"]
|
22 |
+
+ ["contrast", "gaussian_noise", "impulse_noise", "jpeg_compression", "motion_blur", "pixelate", "spatter"]
|
23 |
+
}
|
24 |
+
),
|
25 |
+
)
|
26 |
+
]
|
27 |
+
|
28 |
+
classnames = [
|
29 |
+
"airplane",
|
30 |
+
"airport",
|
31 |
+
"baseball diamond",
|
32 |
+
"basketball court",
|
33 |
+
"beach",
|
34 |
+
"bridge",
|
35 |
+
"chaparral",
|
36 |
+
"church",
|
37 |
+
"circular farmland",
|
38 |
+
"cloud",
|
39 |
+
"commercial area",
|
40 |
+
"dense residential",
|
41 |
+
"desert",
|
42 |
+
"forest",
|
43 |
+
"freeway",
|
44 |
+
"golf course",
|
45 |
+
"ground track field",
|
46 |
+
"harbor",
|
47 |
+
"industrial area",
|
48 |
+
"intersection",
|
49 |
+
"island",
|
50 |
+
"lake",
|
51 |
+
"meadow",
|
52 |
+
"medium residential",
|
53 |
+
"mobile home park",
|
54 |
+
"mountain",
|
55 |
+
"overpass",
|
56 |
+
"palace",
|
57 |
+
"parking lot",
|
58 |
+
"railway",
|
59 |
+
"railway station",
|
60 |
+
"rectangular farmland",
|
61 |
+
"river",
|
62 |
+
"roundabout",
|
63 |
+
"runway",
|
64 |
+
"sea ice",
|
65 |
+
"ship",
|
66 |
+
"snowberg",
|
67 |
+
"sparse residential",
|
68 |
+
"stadium",
|
69 |
+
"storage tank",
|
70 |
+
"tennis court",
|
71 |
+
"terrace",
|
72 |
+
"thermal power station",
|
73 |
+
"wetland",
|
74 |
+
]
|
75 |
+
|
76 |
+
clip_templates = [
|
77 |
+
lambda c: f"satellite imagery of {c}.",
|
78 |
+
lambda c: f"aerial imagery of {c}.",
|
79 |
+
lambda c: f"satellite photo of {c}.",
|
80 |
+
lambda c: f"aerial photo of {c}.",
|
81 |
+
lambda c: f"satellite view of {c}.",
|
82 |
+
lambda c: f"aerial view of {c}.",
|
83 |
+
lambda c: f"satellite imagery of a {c}.",
|
84 |
+
lambda c: f"aerial imagery of a {c}.",
|
85 |
+
lambda c: f"satellite photo of a {c}.",
|
86 |
+
lambda c: f"aerial photo of a {c}.",
|
87 |
+
lambda c: f"satellite view of a {c}.",
|
88 |
+
lambda c: f"aerial view of a {c}.",
|
89 |
+
lambda c: f"satellite imagery of the {c}.",
|
90 |
+
lambda c: f"aerial imagery of the {c}.",
|
91 |
+
lambda c: f"satellite photo of the {c}.",
|
92 |
+
lambda c: f"aerial photo of the {c}.",
|
93 |
+
lambda c: f"satellite view of the {c}.",
|
94 |
+
lambda c: f"aerial view of the {c}.",
|
95 |
+
]
|
96 |
+
|
97 |
+
def _info(self):
|
98 |
+
return datasets.DatasetInfo(
|
99 |
+
description="RESISC45 dataset for image classification.",
|
100 |
+
features=datasets.Features(
|
101 |
+
{
|
102 |
+
"image": datasets.Image(),
|
103 |
+
"label": datasets.ClassLabel(names=self.classnames),
|
104 |
+
}
|
105 |
+
),
|
106 |
+
supervised_keys=("image", "label"),
|
107 |
+
task_templates=[datasets.ImageClassification(image_column="image", label_column="label")],
|
108 |
+
)
|