DOI - Detection of Objects in Images

DOI is an application that searches for images in a provided folder and detects objects in the images.

It uses the YOLOV3 system. It works on GPU with CUDA support, using Darknet For Windows. On CPU, the application uses the DNN module of OpenCV.

This application was built in Python 3.7.

It only supports Windows (for now).

Usage

As stand alone application

This application can be used as a stand alone application. There are two versions:

• Python version: check it out here
• Executable version: check it out here

The executable version does not need instalation.

In Autopsy

In order to integrate DOI with Autopsy, there are two modules:

• DOI Data source ingest module: uses the DOI stand alone application in order to detect objects in images
• DOI Report module: exports a HTML report with the results of the ingest module

Supported image file formats

BMP, DIB, GIF, JPEG, JPEG2000, PBM, PCX, PNG, PPM, SGI, SUN, TGA, TIFF, WEBP and XBM.

Authors

• Mihail Stratan (Instituto Politécnico de Leiria - Portugal)
• Paulo Martinho (Instituto Politécnico de Leiria - Portugal)

Mentors

• Patrício Domingues (Instituto Politécnico de Leiria - Portugal)
• Miguel Frade (Instituto Politécnico de Leiria - Portugal)