nvtnlucie commited on
Commit
e2875d2
·
verified ·
1 Parent(s): 95c57d3

Upload 2 files

Browse files
Files changed (2) hide show
  1. default_crop_v1.md +78 -0
  2. default_crop_v1.pt +3 -0
default_crop_v1.md ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: agpl-3.0
3
+ language:
4
+ - cs
5
+ library_name: ultralytics
6
+ pipeline_tag: object-detection
7
+ tags:
8
+ - cropilot
9
+ - yolo
10
+ - document-layout-analysis
11
+ - document-cropping
12
+ - digitization
13
+ ---
14
+
15
+ # Default Crop v1
16
+
17
+ `default_crop_v1.pt` is a Cropilot model for detecting page crop regions in scanned printed documents. It is the general-purpose default crop model, trained on a mixed corpus of books, magazines, and newspapers.
18
+
19
+ It is published here as a reusable model artifact for use in the [Cropilot](https://github.com/moravianlibrary/cropilot) document-cropping ecosystem.
20
+
21
+ ## License
22
+
23
+ Released under the GNU Affero General Public License v3.0 (`AGPL-3.0`).
24
+
25
+ ## Model File
26
+
27
+ - File: `default_crop_v1.pt`
28
+ - Base model: YOLO11s
29
+ - Task: object-detection (page crop region)
30
+
31
+ ## Intended Use
32
+
33
+ This model is intended for semi-automated processing in a Cropilot workflow:
34
+
35
+ 1. Compressed JPEGs are created from original scans (e.g. TIFF).
36
+ 2. Cropilot runs AI detection using this model.
37
+ 3. A human operator reviews and corrects the result in the Cropilot editor.
38
+ 4. Final crops are applied to the original scans by the production tooling.
39
+
40
+ The model handles both **single-page** and **double-page (spread)** scans, making it a good general default when the document type is mixed or unknown.
41
+
42
+ ## Recommended Integration
43
+
44
+ Typical Cropilot settings for this model:
45
+
46
+ - Crop model: `default_crop`
47
+ - Rotation model: `text`
48
+
49
+ ## Training Data
50
+
51
+ The model was trained on a mixed corpus of digitized printed documents covering both single-page and double-page layouts. Approximate composition:
52
+
53
+ - ~80% books
54
+ - ~15% magazines / periodicals
55
+ - ~5% newspapers
56
+
57
+ The material spans roughly **1950–2010**.
58
+
59
+ ## Limitations
60
+
61
+ - Performance is strongest on recent books; older material, heavily illustrated layouts, or document types outside the training mix may need validation or further fine-tuning.
62
+ - Book covers are detected poorly.
63
+
64
+ ## Example Download
65
+
66
+ ```python
67
+ from huggingface_hub import hf_hub_download
68
+
69
+ path = hf_hub_download(
70
+ repo_id="cropilot-community/default_crop_v1",
71
+ filename="default_crop_v1.pt",
72
+ )
73
+ print(path)
74
+ ```
75
+
76
+ ## Versioning
77
+
78
+ This repository stores **v1** of the default crop model. Newer fine-tuned models should be published as separate `v<n>` repositories so deployments remain reproducible.
default_crop_v1.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:085c797ba97a1145b0a23cb51f8c110d938b9819d3888090e9f998c7444f5b3f
3
+ size 16515394