info@ltutech.com 
+33 1 53 43 01 68 
Linkedin 
Visual search API
Because your trade and specific needs are unique, LTU delivers a suite of SaaS visual search functionalities, open and configurable to recognize, identify and organize your 2D and 3D media.
This SaaS solution gives you access to our multiple proprietary algorithms, and can also integrate third-party algorithms to perform a visual search specifically tailored to your needs.
Based on over 20 years of expertise in image processing technologies, LTU offers a flexible solution while maintaining unmatched performance.
Advanced visual search features
Computer vision without Deep Learning
The core of our technology is based on creating a unique signature of an image or an object. Similarly to an ID, our algorithms extract a wide range of data based on the image or object’s visual characteristics.
This allows it to be immediately operable, with few CPU resources, and to offer unrivaled performances such as a response time <0.3 second on a basis of 1 million images in database, and a daily capacity of 2.5 million requests at a constant response speed of 0.2 second.
Tailored visual search model
We deliver your dedicated visual search model, including the visual processing functionalities specifically required to your needs, in the form of an API key allowing its easy integration into third-party technological environments and applications.
Our support and pre-sales team can also help you for your first steps in the visual recognition world to specify the technological combination best suited to your use case.
Need an extra feature?
Next-level ease of integration
Open API
The solution is available from a Visual Search API that will call the appropriate visual search features. The solution is open and configurable to perform a visual search specific to your needs.
Visual Search API without Deep Learning
The technology is immediately operational, as it doesn’t require any model training based on Deep Learning. The algorithm focuses on visual characteristics (curvature, contrast, etc.), it doesn’t classify an image by typology (landscape, face, animal, etc.)
Comprehensive documentation
Contains all the necessary elements for a straightforward integration via a JSON/REST API. This documentation is in English only.
Dedicated support
A team of experts in image processing, and technical support will help you set up your dedicated object and image recognition API.
Need Change detection also ?
LTU also offers a SaaS solution for comparison and change detection that is highly configurable to operate tailored visual processing to your use case.
Whether you need to set up a level sensitivity to changes, or ignore a background or a reflection, etc. you can adjust the pre and post processing parameters according to your need.
