Integrate Textual and Visual Search using Tensorflow Serving and Elasticsearch
This demo is created from visual-search-v2.0, which includes color filter, numeric filter and text search. We are going to update this git repo (visual-search-v1.0) soon.
Create a visual search platform in one step:
# git clone [email protected]:gyang274/visual-search.git
git clone https://github.com/gyang274/visual-search.git
cd visual-search
docker-compose up
Open your favorite browser with https://127.0.0.1/#/, try some URL and local file with visual search, e.g., https://www.potterybarn.com/pbimgs/ab/images/dp/wcm/201824/0483/trieste-dining-chair-o.jpg
The docker image contains 1080 furniture images by default. You can add your own images using the Manage page, or using the scripts provided by this repository.
-
How to add images using the
Managepage. -
How to add images in bulk using docker cp and python scripts.