<img src="images/box-dev-logo-clip.png" alt= “box-dev-logo” style="margin-left:-10px;" width=40%;>
These are community created Box Skills samples for processing image files on Box. By clicking on each sub-folder, you can see a demo image of the end product, as it would look in your Box file preview.
Acuant AssureID Government ID Data Extraction - Uses Acuant AssureID to classify and automatically extract metadata from a government-issued ID, such as name, numbers, address, etc, and add them as Skills Metadata Cards to the respective image files on Box.
Amazon Rekognition Labels Detection - Uses the Amazon Rekognition API to automatically extract labels, which are objects, events, or concepts that are present in the image provided and add them to the respective image files as Skills Metadata Cards.
EXIF/XMP MetaInfo Extraction - Uses the Javascript rewrite of the decade long opensource developed Exiftool to read dozens of MetaInfo from files. Works for all Image, Audio, Video files.
Google Product Search recognition - Uses the Google Cloud Vision Product Search to compare images in Box with the images in a pretrained product catalog.
Google Image Text and Topics - Uses the Google Cloud Vision API to extract image texts and topics, using generic Google Image Search ML model for identification.
Hive Predict face recognition - Uses the Hive Predict API to automatically recognize faces in images and assign these faces as Skills Metadata Cards to the respective images in Box.
Microsoft Image Text and Topics - Uses the Microsoft Vision API to extract image texts and topics, using generic Bing Image Search ML model for identification.
Box Skills are web applications configured with Box Platform that performs custom processing for files uploaded to Box. Typically they link to a machine learning service that does the processing for the files. Your choice of machine learning service, in-house or external, would depend on your business case or that of your customers on Box. However, the audio, video, image and document sample Skills repositories in Box Community can provide some guidance or inspiration on what you can built upon or deployed as-is.
Visit the Official Box Skills Developer Documentation for complete information on Box Skills, the kind of Skill Metadata Cards that you can create to show on Box Preview, as well as instructions on configuring your Skill with Box.
The file formats supported by your skill depend on the files supported by Box and by the selected machine learning service.
Box supports the following image formats for direct previewing using Box Preview. The various machine learning providers often support a different set of file formats completely.
You can expand the effective set of file formats supported by your machine learning provider by using Box's BasicFormat functionality in the Skills Kit. This automatically converts some file formats to more commonly used formats for you to use in your skill.
The Github Repository for Box Skills Kit Library is our official toolkit for writing custom Box Skills in Node.js. It minimizes the client side code to Box Files and Skills-Invocations APIs to a few lines and provides other utility functions to make developing your code very simple. It has the Skills-kit Library and API Documentation and Boilerplate Skills that you can quickly deploy and expand on, when developing a new Skill.
In developing your custom Box Skill, you would need to deploy it somewhere. Have a look at our Quick Start Deployment Instructions to learn how to deploy the boilerplate Skills to any of your preferred cloud server providers or on your own server environment. Additionally, each of the sample Skills in this repository may give extra or alternative deployment instructions, that you can use.
This project is licensed under the Apache 2.0 License