UHMC/object-analysis

object-analysis

Introduction

This project’s aim is to conceptualize and implement an open-source, simple, and extensible framework for continuous image analysis using the TensorFlow Object Detection API.

System Requirements

• Ubuntu Linux ver. 16.04

Dependencies

The setup script should take care of any dependencies automatically; however, it should be noted that these packages will be installed on your system if all goes well:

• mysql-server
• git (if not installed already)
• protobuf-compiler
• python-pip
  • tensorflow
  • pillow
  • lxml
  • jupyter
  • matplotlib
  • ipython
  • mysql-connector=2.1.4
  • jsonpickle
  • exifread

File Hierarchy

The setup script will create the following folder structure used by the database and the ongoing processing scripts:

• /srv/ObjectDB
  • /EXIF
  • /odapi_output
  • /unprocessed
  • /processed

All folders are owned by the mysql usergroup, and the current user at setup time will be added to the mysql group.

Demo Instructions

1. Install Git if necessary:
   $ sudo apt install git -y
2. Clone the repository:
   $ git clone https://github.com/UHMC/object-analysis.git
3. Change directory into the project folder:
   $ cd object-analysis
4. Run the setup script:
   $ ./setup_ubuntu1604.sh
5. Add demo images to unprocessed folder:
   $ cp Resources/demo-images/* /srv/ObjectDB/unprocessed/
6. Wait.

Visual Diagram

Screenshots

Credits

• TensorFlow Object Detection API
  • "Speed/accuracy trade-offs for modern convolutional object detectors." Huang J, Rathod V, Sun C, Zhu M, Korattikara A, Fathi A, Fischer I, Wojna Z, Song Y, Guadarrama S, Murphy K, CVPR 2017
• Mario Canul - Undergraduate Cybersecurity Researcher
• Saxon Knight - Undergraduate Cybersecurity Researcher