Image Annotation Module - ShehanWeerakkody/OpenDF GitHub Wiki

Nowadays, digital forensic tools are often used to investigate cyber crimes, which are becoming more and more frequent. OpenDF is a tool for forensic analysis of disk data. By having an image description module that automatically recognises objects from the large amount of images that are available in a disk, it saves manual work for an investigator. Basically, there is a need to maintain all the objects detected in an image in a Database and make it searchable using keywords.

This REST API will take the image uuid as the input and returns a JSON with predicted values. This contains code and weights files for the following pre-trained models.

• VGG16
• VGG19
• ResNet50

Moreover we include Google Cloud Vision API for this module. This has a huge amount of performance improvement. The image annotation results will return within 3 seconds. If the investigation organisation prefers to use Google Cloud Vision API then they should be able to change configs for this.

Following is an example for API usage.