Update Populus_Stomatal_Images_Datasets.py
Browse files
Populus_Stomatal_Images_Datasets.py
CHANGED
|
@@ -85,7 +85,8 @@ class NewDataset(datasets.GeneratorBasedBuilder):
|
|
| 85 |
"image_id": datasets.Value("string"),
|
| 86 |
"species": datasets.Value("string"),
|
| 87 |
"scientific_name": datasets.Value("string"),
|
| 88 |
-
"pics_array": datasets.
|
|
|
|
| 89 |
"image_resolution": {
|
| 90 |
"width": datasets.Value("int32"),
|
| 91 |
"height": datasets.Value("int32"),
|
|
@@ -179,8 +180,9 @@ class NewDataset(datasets.GeneratorBasedBuilder):
|
|
| 179 |
|
| 180 |
pics_array = None
|
| 181 |
with Image.open(image_path) as img:
|
| 182 |
-
pics_array = np.array(img)
|
| 183 |
-
|
|
|
|
| 184 |
annotations = self._parse_yolo_labels(label_path, width, height)
|
| 185 |
|
| 186 |
# Yield the dataset example
|
|
|
|
| 85 |
"image_id": datasets.Value("string"),
|
| 86 |
"species": datasets.Value("string"),
|
| 87 |
"scientific_name": datasets.Value("string"),
|
| 88 |
+
"pics_array": datasets.Image()
|
| 89 |
+
# datasets.Array3D(dtype="uint8", shape=(3,768, 1024)), # Assuming images are RGB with shape 768x1024
|
| 90 |
"image_resolution": {
|
| 91 |
"width": datasets.Value("int32"),
|
| 92 |
"height": datasets.Value("int32"),
|
|
|
|
| 180 |
|
| 181 |
pics_array = None
|
| 182 |
with Image.open(image_path) as img:
|
| 183 |
+
pics_array = np.array(img)# Convert the PIL image to a numpy array and then to a list
|
| 184 |
+
print(pics_array.shape)
|
| 185 |
+
|
| 186 |
annotations = self._parse_yolo_labels(label_path, width, height)
|
| 187 |
|
| 188 |
# Yield the dataset example
|