iarata commited on
Commit
87b981c
1 Parent(s): 3ef530c

Initial commit

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
PHCR_DB25_2775.zip ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ed5db73f5062bfb80c10c0d541fbbc1e39423d92ccab7120f765589724318f4
3
+ size 8214798
README.md ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - fa
4
+ pretty_name: Persian Historical Documents Handwritten Characters
5
+ size_categories:
6
+ - 1K<n<10K
7
+ tags:
8
+ - ocr
9
+ - character-recognition
10
+ - persian
11
+ - historical
12
+ - handwritten
13
+ - nastaliq
14
+ - character
15
+ ---
16
+
17
+ # Persian Historical Documents Handwritten Characters
18
+
19
+ ## Dataset Description
20
+
21
+ - **Repository:** https://github.com/iarata/persian-docs-ocr
22
+ - **Paper:** https://doi.org/10.1007/978-3-031-53969-5_20
23
+ - **Point of Contact:** hajebrahimi.research [at] gmail [dot] com
24
+
25
+ ### Summary
26
+
27
+ This dataset contains pre-processed images of Persian characters' contextual forms (except letter گ) from 5 handwritten Persian historical books written in Nastaliq script. The dataset contains 2775 images of 111 classes. The images are in TIFF format and have a resolution of 72 dpi. The images are in black and white and have a size of 395 × 395 pixels.
28
+
29
+ ### Languages
30
+
31
+ Persian
32
+
33
+ ![Sample view of the dataset](dataset-sample-view.png)
34
+
35
+ ## Dataset Structure
36
+
37
+ The dataset is structured as follows:
38
+
39
+ ```
40
+ ├── data
41
+ │ ├── 06a9_01.tif
42
+ │ ├── 06a9_02.tif
43
+ │ ├── 06a9_03.tif
44
+ │ ├── 06a9_04.tif
45
+ │ ├── 06a9_05.tif
46
+ │ ├── ...
47
+ │ ├── 06a9_25.tif
48
+ │ │
49
+ │ ├── 06cc_01.tif
50
+ │ ├── 06cc_02.tif
51
+ │ ├── 06cc_03.tif
52
+ │ ├── 06cc_04.tif
53
+ │ ├── 06cc_05.tif
54
+ │ ├── ...
55
+ │ ├── 06cc_25.tif
56
+ │ ├── ...
57
+ ```
58
+
59
+ The naming of each image indicates the UTF-16 hexadecimal code ([Hex to String Decoder](https://dencode.com/en/string/hex)) of a character's contextual form followed by the number of the image. In the numbering, every 5 images are from a new book. The contextual form of every character is treated as a separate class resulting in 111 classes.
60
+
61
+ ## Dataset Creation
62
+
63
+ For building this dataset 5 historical Persian books from the [Library of Congress](loc.gov)
64
+
65
+ ### Source Data
66
+
67
+
68
+ The data was collected from 5 historical Persian books from the [Library of Congress](loc.gov). The books are as follows:
69
+
70
+ - [Shah-nameh by Firdausi](https://www.loc.gov/item/2012498868/)
71
+ - [Dīvān](https://www.loc.gov/item/2015481730/)
72
+ - [Kitāb-i Rūmī al-Mawlawī](https://www.loc.gov/item/2016397707)
73
+ - [Gulistān](https://www.loc.gov/item/2017406684/)
74
+ - [Qajar-era poetry](https://www.loc.gov/item/2017498320/)
75
+
76
+ The images were pre-processed using the following steps:
77
+
78
+ Images were first normalized to reduce noise from the background of the characters. The normalized image is then converted to a single-channel grayscale image. Following that, image thresholding is applied to the grayscale image to remove the characters' background. The thresholded image is binarized so that the pixel values greater than 0 become 255 (white), and pixels with a value of 0 (black) remain unchanged. Finally, the binarized image is inversed.
79
+
80
+ ### Annotations
81
+
82
+ Before pre-processing the images the characters were cropped from the books and were saved with their UTF-16 hexadecimal code plus the number of the image (e.g. 06a9_01.tif).
83
+
84
+ #### Annotators:
85
+ - [Hajebrahimi Alireza](https://www.linkedin.com/in/alireza-hajebrahimi/)
86
+ - [Hajebrahimi Reyhaneh](https://www.linkedin.com/in/reyhaneh-hajebrahimi-2565451a0/)
87
+
88
+
89
+ ### Citation Information
90
+
91
+ Hajebrahimi, A., Santoso, M.E., Kovacs, M., Kryssanov, V.V. (2024). Few-Shot Learning for Character Recognition in Persian Historical Documents. In: Nicosia, G., Ojha, V., La Malfa, E., La Malfa, G., Pardalos, P.M., Umeton, R. (eds) Machine Learning, Optimization, and Data Science. LOD 2023. Lecture Notes in Computer Science, vol 14505. Springer, Cham. https://doi.org/10.1007/978-3-031-53969-5_20
92
+
93
+ **BibTeX:**
94
+
95
+ ```bibtex
96
+ @InProceedings{10.1007/978-3-031-53969-5_20,
97
+ author="Hajebrahimi, Alireza
98
+ and Santoso, Michael Evan
99
+ and Kovacs, Mate
100
+ and Kryssanov, Victor V.",
101
+ editor="Nicosia, Giuseppe
102
+ and Ojha, Varun
103
+ and La Malfa, Emanuele
104
+ and La Malfa, Gabriele
105
+ and Pardalos, Panos M.
106
+ and Umeton, Renato",
107
+ title="Few-Shot Learning for Character Recognition in Persian Historical Documents",
108
+ booktitle="Machine Learning, Optimization, and Data Science",
109
+ year="2024",
110
+ publisher="Springer Nature Switzerland",
111
+ address="Cham",
112
+ pages="259--273",
113
+ abstract="Digitizing historical documents is crucial for the preservation of cultural heritage. The digitization of documents written in Perso-Arabic scripts, however, presents multiple challenges. The Nastaliq calligraphy can be difficult to read even for a native speaker, and the four contextual forms of alphabet letters pose a complex task to current optical character recognition systems. To address these challenges, the presented study develops an approach for character recognition in Persian historical documents using few-shot learning with Siamese Neural Networks. A small, novel dataset is created from Persian historical documents for training and testing purposes. Experiments on the dataset resulted in a 94.75{\%} testing accuracy for the few-shot learning task, and a 67{\%} character recognition accuracy was observed on unseen documents for 111 distinct character classes.",
114
+ isbn="978-3-031-53969-5"
115
+ }
116
+ ```
data/0627_01.tif ADDED
data/0627_02.tif ADDED
data/0627_03.tif ADDED
data/0627_04.tif ADDED
data/0627_05.tif ADDED
data/0627_06.tif ADDED
data/0627_07.tif ADDED
data/0627_08.tif ADDED
data/0627_09.tif ADDED
data/0627_10.tif ADDED
data/0627_11.tif ADDED
data/0627_12.tif ADDED
data/0627_13.tif ADDED
data/0627_14.tif ADDED
data/0627_15.tif ADDED
data/0627_16.tif ADDED
data/0627_17.tif ADDED
data/0627_18.tif ADDED
data/0627_19.tif ADDED
data/0627_20.tif ADDED
data/0627_21.tif ADDED
data/0627_22.tif ADDED
data/0627_23.tif ADDED
data/0627_24.tif ADDED
data/0627_25.tif ADDED
data/0628_01.tif ADDED
data/0628_02.tif ADDED
data/0628_03.tif ADDED
data/0628_04.tif ADDED
data/0628_05.tif ADDED
data/0628_06.tif ADDED
data/0628_07.tif ADDED
data/0628_08.tif ADDED
data/0628_09.tif ADDED
data/0628_10.tif ADDED
data/0628_11.tif ADDED
data/0628_12.tif ADDED
data/0628_13.tif ADDED
data/0628_14.tif ADDED
data/0628_15.tif ADDED
data/0628_16.tif ADDED
data/0628_17.tif ADDED
data/0628_18.tif ADDED
data/0628_19.tif ADDED
data/0628_20.tif ADDED
data/0628_21.tif ADDED
data/0628_22.tif ADDED
data/0628_23.tif ADDED