File size: 4,115 Bytes
af85382
 
0fe2c4e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
af85382
e69cc11
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
---
license: cc-by-nc-4.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: val
    path: data/val-*
dataset_info:
  features:
  - name: image_id
    dtype: string
  - name: image
    dtype: image
  - name: width
    dtype: int64
  - name: height
    dtype: int64
  - name: objects
    struct:
    - name: bbox
      sequence:
        sequence: float64
    - name: category
      sequence: int64
    - name: id
      sequence: 'null'
  splits:
  - name: train
    num_bytes: 19639133.0
    num_examples: 79
  - name: val
    num_bytes: 4967295.0
    num_examples: 21
  download_size: 24112875
  dataset_size: 24606428.0
---

---
license: cc-by-nc-4.0
task_categories:
- object-detection
language:
- la
tags:
- object detection
- critical edition
- yolo
size_categories:
- n<1K
---

# MGH Layout Detection Dataset

## Dataset Description

### General Description

This dataset consists of scans from the MGH critical edition of Alcuin's letters, which were first edited by Ernestus Duemmler in 1895. The digital scans were sourced from the DMGH's repository, which can be accessed [here](https://www.dmgh.de/mgh_epp_4). The scans were annotated using CVAT, marking out two classes: the title of a letter and the body of the letter.

### Why was this dataset created?

The primary motivation behind the creation of this dataset was to enhance the downstream task of OCR. OCR often returns errors due to interferences like marginalia and footnotes present in the scanned pages. By having accurate annotations for the title and body of the letters, users can efficiently isolate the main content of the letters and possibly achieve better OCR results.

Future plans for this dataset include expanding the annotations to encompass footnotes and marginalia, thus further refining the demarcation between the main content and supplementary notes.

### Classes

Currently, the dataset has two annotated classes:
- Title of the letter
- Body of the letter

Planned future additions include:
- Footnotes
- Marginalia

## Sample Annotation

![sample_annotation](sample_annotation.JPG)

## Biographical Information

### About Alcuin

Alcuin of York (c. 735 – 804 AD) was an English scholar, clergyman, poet, and teacher. He was born in York and became a leading figure in the so-called "Carolingian renaissance." Alcuin made significant contributions to the educational and religious reforms initiated by Charlemagne, emphasizing the importance of classical studies.

### About Alcuin's Letters

Alcuin's letters provide a crucial insight into the Carolingian world, highlighting the intellectual and religious discourse of the time. They serve as invaluable resources for understanding the interactions between some of the important figures of Charlemagne's court, the challenges they faced, and the solutions they proposed. The letters also offer a window into Alcuin's own thoughts, his relationships with peers and, most importantly, his students, and his role as an advisor to Charlemagne.

## Dataset and Annotation Details

### Annotation Process

The scans of Alcuin's letters were annotated manually using the CVAT tool. The primary focus was to delineate the titles and bodies of the letters. This clear demarcation aids in improving the precision of OCR tools by allowing them to target specific regions in the scanned pages.

### Dataset Limitations

As the dataset currently focuses only on titles and bodies of the letters, it may not fully address the challenges posed by marginalia and footnotes in OCR tasks. However, the planned expansion to include these classes will provide a more comprehensive solution.

### Usage

Given the non-commercial restriction associated with the source scans, users of this dataset should be mindful of the [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)](https://creativecommons.org/licenses/by-nc-sa/4.0/) license under which it is distributed.

## Additional Information

For more details on the dataset and to access the digital scans, visit the DMGH repository link provided above.