# Elasticsearch and Machine Learning

**URL:** <https://discuss.elastic.co/t/elasticsearch-and-machine-learning/59896>\
**Category:** Elasticsearch\
**Created:** [September 6, 2016, 4:21pm UTC](https://discuss.elastic.co/t/elasticsearch-and-machine-learning/59896 "2016-09-06T16:21:57Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![kmk3173](https://avatars.discourse-cdn.com/v4/letter/k/8baadc/32.png) [@kmk3173](https://discuss.elastic.co/u/kmk3173)\
**Post date:** [September 6, 2016, 4:21pm UTC](https://discuss.elastic.co/t/elasticsearch-and-machine-learning/59896/1 "2016-09-06T16:21:57Z")

</div>

Has anyone successfully integrated LiVRE and elasticsearch? I'm looking for an open-source plugin or possible solution for a video content based image retrieval system using elasticsearch.

---

<div class="post-metadata">

**Author:** ![polyfractal](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/polyfractal/32/48162_2.png) [@polyfractal](https://discuss.elastic.co/u/polyfractal)\
**Post date:** [September 6, 2016, 5:12pm UTC](https://discuss.elastic.co/t/elasticsearch-and-machine-learning/59896/2 "2016-09-06T17:12:13Z")

</div>

I'm not familiar with LiVRE... did you mean LIRE? There are some published articles and plugins using LIRE to do image search in ES:

- [https://www.elastic.co/blog/found-getting-started-with-lire-and-elasticsearch](https://www.elastic.co/blog/found-getting-started-with-lire-and-elasticsearch)
- [https://github.com/kzwang/elasticsearch-image](https://github.com/kzwang/elasticsearch-image)

There's also been some more recent articles/tutorials on using DeepDetect to do image search in ES:

- [https://www.elastic.co/blog/categorizing-images-with-deep-learning-into-elasticsearch](https://www.elastic.co/blog/categorizing-images-with-deep-learning-into-elasticsearch)
- [http://www.deepdetect.com/tutorials/es-image-classifier/](http://www.deepdetect.com/tutorials/es-image-classifier/)

And then there are more manual methods, such as using perceptual hashing or just extracting color palette compositions:

- [https://stackoverflow.com/questions/32785803/similar-image-search-by-phash-distance-in-elasticsearch](https://stackoverflow.com/questions/32785803/similar-image-search-by-phash-distance-in-elasticsearch)
- [http://blog.sandeepchivukula.com/posts/2016/03/06/photo-search/](http://blog.sandeepchivukula.com/posts/2016/03/06/photo-search/)
- [https://sujitpal.blogspot.com/2016/05/elasticsearch-based-image-search-using.html](https://sujitpal.blogspot.com/2016/05/elasticsearch-based-image-search-using.html)
- [https://www.elastic.co/elasticon/conf/2016/sf/opensource-connections-the-ghost-in-the-search-machine](https://www.elastic.co/elasticon/conf/2016/sf/opensource-connections-the-ghost-in-the-search-machine)

They all revolve around the same idea though: take an image (or still frame from a video) and produce a set of tokens which can be indexed. These tokens can correspond to colors, contrast, scale, orientation, etc. And then at search time, you tokenize the query image the same way and let lucene find the matching tokens / docs (images).

---

<div class="post-metadata">

**Author:** ![kmk3173](https://avatars.discourse-cdn.com/v4/letter/k/8baadc/32.png) [@kmk3173](https://discuss.elastic.co/u/kmk3173)\
**Post date:** [September 6, 2016, 5:19pm UTC](https://discuss.elastic.co/t/elasticsearch-and-machine-learning/59896/3 "2016-09-06T17:19:20Z")

</div>

Its similar to LiRE but based on videos.

[http://nospotfer.github.io/livre/](http://nospotfer.github.io/livre/)

---

<div class="post-metadata">

**Author:** ![polyfractal](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/polyfractal/32/48162_2.png) [@polyfractal](https://discuss.elastic.co/u/polyfractal)\
**Post date:** [September 6, 2016, 5:47pm UTC](https://discuss.elastic.co/t/elasticsearch-and-machine-learning/59896/4 "2016-09-06T17:47:52Z")

</div>

Ah, I'm not aware of any plugins/tutorials that use that. Just static images, so you'd have to pull out keyframes yourself and use existing methods. Or start to integrate LIvRE yourself in a plugin.

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [July 5, 2017, 10:22pm UTC](https://discuss.elastic.co/t/elasticsearch-and-machine-learning/59896/5 "2017-07-05T22:22:14Z")

</div>


