davanstrien HF staff commited on
Commit
76a76cc
1 Parent(s): c7d1289

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +4 -0
app.py CHANGED
@@ -16,7 +16,11 @@ DESCRIPTION = """
16
 
17
  [IIIF](https://iiif.io/get-started/) is a standard for sharing images that many libraries, archives and museums use. One part of the IIIF standard is the [presentation API](https://iiif.io/api/presentation/3.0/) which defines a way of representing multiple images which make up a digitised physical item such as a book as a JSON document. This app allows you to pass the URL for one of these manifests and return all of the book pages containing illustrations.
18
 
 
 
19
  You can read more about how we created the model and training data for this project in this [GitHub repository](https://github.com/davanstrien/ImageIN/)
 
 
20
  """
21
 
22
  HF_MODEL_PATH = (
 
16
 
17
  [IIIF](https://iiif.io/get-started/) is a standard for sharing images that many libraries, archives and museums use. One part of the IIIF standard is the [presentation API](https://iiif.io/api/presentation/3.0/) which defines a way of representing multiple images which make up a digitised physical item such as a book as a JSON document. This app allows you to pass the URL for one of these manifests and return all of the book pages containing illustrations.
18
 
19
+ For example, [Harvard Library Digital Collections](https://library.harvard.edu/digital-collections) has many digitised books. Many of these books have IIIF manifests available. If we take [this](https://digitalcollections.library.harvard.edu/catalog/990052690260203941) book as an example, we can find a [link](https://iiif.lib.harvard.edu/manifests/drs:49040527) to it's manifest. If we pass this manifest to the app, we should get all the pages from that book containing illustrations. This is done by grabbing all the individual page URLs from the book (using the manifest) and passing them through the image classification model we created as part of this project.
20
+
21
  You can read more about how we created the model and training data for this project in this [GitHub repository](https://github.com/davanstrien/ImageIN/)
22
+
23
+
24
  """
25
 
26
  HF_MODEL_PATH = (