Maurizio is a Tech Lead at Alfresco with a background in BPM, banking, and insurance. He is leading the Process Workspace App, an Angular app based on Activiti 7 Cloud with a focus on the CI/CD. In his spare time, Maurizio likes to learn more about TensorFlow.js as well as Google Firebase.
Searching is one of the most popular activity users conduct on web applications, which yields any kind of content, such as products, images, or news.
In the case of applications handling images, the search procedure typically involves matching only existing metadata, e.g., name, date, and size. However, users rarely know this information due to the large amount of data nowadays available. In such a circumstance, machine learning can be helpful.
We propose a method for integrating web applications (Angular) with machine learning (TensorFlowJs), which facilitates user searching in the following ways. First, it enriches existing image metadata with machine learning by adding further information provided by object detector models. Second, it enables users to search any items by image content rather than existing metadata. Specifically, this workshop will show you how to a) create an angular app utilizing angular cli, b) save permanently images on google storage, c) make use of machine learning models for object recognition and detection, d) save predicted metadata on Google fire cloud, e) finally, search any items by image content.