SentiSight web service is an interactive platform for developing AI-based image recognition applications. This platform equips you with tools that give guidance through the development process. With the technical background provided, the process will be more intuitive and fast.


Skipping the technical details, below are the steps of building a model for successful image recognition that you will use in your application.
1) Collect and label a large enough number of images providing information on their content. Labeling the information depends on the type of model  you choose: image classification, object detection, and instance segmentation).

2) Train a model (i.e., develop an algorithm) that learns from your data. The algorithm will iterate through all the images until it is able to predict the context of those ‘not seen’ before. You will also need to test the model to find out what still has to be improved.
3) Use your model online. Integrating it into your system would be a matter of a Custom Project.

The tools we propose for the interactive web platform SentiSight is the outcome of our 28 years experience in algorithm engineering. They were designed to support faster image labeling and interactive model training. You will not have to worry about writing many lines of code, and this will save a lot of your time in all the stages of development including building a model, testing it and exploring the results.