Object Identification
Identification of relevant information in large amounts of data in pictures and videos is not only time-consuming but often almost impossible for the human eye. In addition, human and time resources are limited. These conditions quickly present you with real challenges: Is relevant information included and how do you get timely results? Automated support can help.
In one specific case, the securing of mobile devices and data carriers, there were more than 40,000 images and several hundred videos in digital evidence objects. Our solution offers – with the help of neural networks – the automated identification and marking of objects such as vehicles, weapons or symbols in images and videos. This made a fast and prioritized inspection possible.
FUNCTIONALITIES:
- Identification of objects in
- Images (jpg, jpeg, png, bmp, etc.)
- Videos (mpeg, avi, mp4, mov, etc.)
- In streams
- Highlighting of found objects on pictures and in videos
- Extraction of existing EXIF data
- Enrichment with metadata, for example vehicle brand & model and associated confidence value
- Export to CSV or JSON
- User based retraining of the individual neural networks
Available neural networks for object recognition:
- Vehicles
- Weapons
- Symbols
- Further networks in preparation and on request