Several new tools developed at the Griffeye Hackathon: Griffhack
A dedicated team of developers and end users brought their creativity, ideas and knowledge together to develop some great new tools at the Griffeye Hackathon. They are now released.
A couple of weeks ago, we held a two-day Hackathon called Griffhack at our office in Gothenburg, Sweden. Griffhack was a get-together for users, developers and programmers with the shared mission to come up with new software technologies to extend and innovate the Analyze platform even further. It was a great way to combine creative ideas and skills, get input from users and highlight the team core strengths.
Always with the user in mind
We also invited International Law Enforcement partners to bring ideas to develop tools that would really benefit our users. They gave us really valuable input on a lot of do’s and don’ts about what would work. Based on their input and feedback, and the creative minds of our developers, we created some really great stuff. And by the end of Griffhack we had three nearly finished plugins: the object classifier, XMP data extractor and video face recognition. All of which have been modified and improved to be implemented in the 18.01 release.
The three new Griffeye plugins
1. Object classifier
The object classifier (also called Griffeye Brain Objects) is a plugin that classifies images in different categories and subcategories based on the most significant object in the image. You can also add or remove objects from the search to reduce irrelevant data. The plugin can be used in combination with, for example, Face Recognition for further specific filtering. This first version is trained on single-object images. Multi-object image support will make it into a future release.
2. XMP data extractor
XMP (Extensible Metadata Platform) is the creation, processing and interchange of metadata for all kinds of resources. The Griffeye XMP Grouping is a plugin that extracts XMP metadata. The plugin goes through all files in batches and locates the Document IDs to find if there are any relations to other files within a case. This first version is a proof of concept. If it works well, more can be done to extract information during import – and possibly find other XMP data fields that is of use to investigators.
3. Video face recognition
In the previous Griffeye Analyze 18.0 release, we launched the new high-performance Face Recognition app that enables users to quickly find images in a case that contains similar faces to a suspect or victim.It will now also be possible to detect and recognize faces in videos. The video face detection is done automatically at import and it is also possible find and search for specific suspects or victims in videos across your case.
READ MORE: Release of Analyze 18.1