Reference implementation on building a local image search engine that uses Amazon Rekognition image labels to create an inverted index.
(Tested on Mac OSX)
-
pip install pillow
-
pip install wordcloud
-
Create a directory called 'photos' and add some images in it.
-
Run the indexer:
$ python2.7 indexer.py -
A new file called web/photo_tags.png is created with the tags of the photos.
This code sample is licensed under the MIT-0 License.