ducto489 commited on
Commit
bb93659
·
verified ·
1 Parent(s): 341eb1f

Create README.md

Browse files
Files changed (1) hide show
  1. README.md +64 -0
README.md ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ language:
4
+ - en
5
+ - vi
6
+ tags:
7
+ - ocr
8
+ - optical-character-recognition
9
+ - text-recognition
10
+ - image-text
11
+ pretty_name: Combined OCR Datasets
12
+ size_categories:
13
+ - 10M<n<100M
14
+ task_categories:
15
+ - image-text-recognition
16
+ task_ids:
17
+ - optical-character-recognition
18
+ ---
19
+
20
+ # Combined OCR Dataset for Text Recognition
21
+
22
+ ## Dataset Description
23
+
24
+ This is a large-scale dataset (~11M training, ~0.9M validation images) for Optical Character Recognition (OCR), aggregated from several common benchmarks and sources (see Sources below). It includes scene text, handwritten text, and synthetic images with corresponding text labels.
25
+
26
+ **Training Code:** [https://github.com/AnyGlow/lib_ocr](https://github.com/AnyGlow/lib_ocr)
27
+
28
+ ## Dataset Structure
29
+
30
+ ```
31
+ ./data/
32
+ ├── train/
33
+ │ ├── images/*.jpg # Training images
34
+ │ └── tgt.csv # Training filenames and labels
35
+ └── val/
36
+ ├── images/*.jpg # Validation images
37
+ └── tgt.csv # Validation filenames and labels
38
+ ```
39
+ * `tgt.csv`: Comma-separated `image_name,label`. Labels are pre-normalized (see below).
40
+
41
+ ## Data Details
42
+
43
+ * **Sources:** Aggregated from [vietocr](https://github.com/pbcquoc/vietocr), [Deep Text Rec. Benchmark](https://github.com/clovaai/deep-text-recognition-benchmark), [Synth90k](https://www.robots.ox.ac.uk/~vgg/data/text/), and [Cinnamon AI Handwritten](https://www.kaggle.com/datasets/hariwh0/cinnamon-ai-handwritten-addresses).
44
+ * **Normalization:** Text labels have undergone character normalization (e.g., smart quotes to `"` , various dashes to `-`, accented chars simplified like `Ö`->`O`). This simplifies the vocabulary. See the [training repo README](https://github.com/AnyGlow/lib_ocr#data-preprocessing-character-normalization) for the full list.
45
+ * **Language:** Predominantly English. Vietnamese samples (1.76%) are included but limited.
46
+ * **Licensing:** Combines multiple sources. **Users must verify compliance with original dataset licenses** before use.
47
+
48
+ ## Citation
49
+
50
+ Please cite the original source datasets if you use this data. Citing this aggregation is also appreciated:
51
+
52
+ ```bibtex
53
+ @misc{ducto489_ocr_datasets,
54
+ author = {Nguyễn Minh Đức},
55
+ title = {Combined OCR Dataset},
56
+ year = {2024}, # Update year as needed
57
+ howpublished = {Hugging Face Datasets},
58
+ url = {https://huggingface.co/datasets/ducto489/ocr_datasets}
59
+ }
60
+ ```
61
+
62
+ ## Dataset Curators
63
+
64
+ Aggregated by [ducto489](https://github.com/ducto489/) / Nguyễn Minh Đức.