riotu-lab commited on
Commit
4ee7e74
·
verified ·
1 Parent(s): 9d58a8b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +55 -1
README.md CHANGED
@@ -19,4 +19,58 @@ features:
19
  dtype: string
20
  tags:
21
  - dataset
22
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
19
  dtype: string
20
  tags:
21
  - dataset
22
+ ---
23
+
24
+ ### Dataset Description
25
+
26
+ This dataset is designed for training and evaluating Optical Character Recognition (OCR) models
27
+ for Arabic text. It is an extension of an open-source dataset and includes text rendered in multiple Arabic fonts (Amiri, Sakkal Majalla, Arial, Calibri and Scheherazade New).
28
+ The dataset simulates real-world book layouts to enhance OCR accuracy.
29
+
30
+ ### Dataset Structure
31
+ The dataset is divided into five splits based on font name (Sakkal_Majalla, Amiri, Arial, Calibri, and Scheherazade_New).
32
+ Each split contains data specific to a single font. Within each split, the following attributes are present:
33
+ - **image_name**: Unique identifier for each image.
34
+ - **chunk**: The text content associated with the image.
35
+
36
+ - **font_name**: The font used in text rendering.
37
+
38
+ - **image_base64**: Base64-encoded image representation.
39
+
40
+ ### How to Use
41
+
42
+ ```python
43
+ from datasets import load_dataset
44
+ import base64
45
+ from io import BytesIO
46
+ from PIL import Image
47
+ # Load dataset with streaming enabled
48
+ ds = load_dataset("xya22er/text_to_image", streaming=True)
49
+ print(ds)
50
+
51
+
52
+
53
+
54
+ # Load the dataset
55
+
56
+ # Iterate over a specific font dataset (e.g., Amiri)
57
+ for sample in ds["Amiri"]:
58
+ image_name = sample["image_name"]
59
+ chunk = sample["chunk"] # Arabic text transcription
60
+ font_name = sample["font_name"]
61
+
62
+ # Decode Base64 image
63
+ image_data = base64.b64decode(sample["image_base64"])
64
+ image = Image.open(BytesIO(image_data))
65
+
66
+ # Show the image (optional)
67
+ image.show()
68
+
69
+ # Print the details
70
+ print(f"Image Name: {image_name}")
71
+ print(f"Font Name: {font_name}")
72
+ print(f"Text Chunk: {chunk}")
73
+
74
+ # Break after one sample for testing
75
+ break
76
+ ```