Springe zum Hauptinhalt
Technische Informatik
Technische Informatik
Technische Informatik 

IFE Tool

IFE, which stands for Intelligent Frame Extractor, is a software application that was developed by the Chair of Computer Engineering to support the different projects under study at the IFC (Indoor Flight Center).

IFE is a data processing tool that is used for the processing of video frames using different object detection algorithms. The results of the processing are saved by the mean of data serialization techniques

A range of features are implemented within IFE to support the visualization of the videos and the results of the processing

The IFE application has a modular architecture to support reusability. New object detection algorithms can be integrated, along with their relative features, as an extension module. This allows the application to adapt to different project applications.

  • Data Analysis::
    • Data serialization
      • Logging of detection results
      • Support for reusability and portability
    • Report generation
      • Data analysis based on the serialization files
      • Generation of an inspection report (support for pdf format)
    • Heatmap generation
      • A heatmap-inspired feature that allows to highlight the important parts (frames with detected object) within a video
      • Support for multiple objects
  • Image Processing: :
    • Modular architecture
      • C++/python support
      • Modular integration of Image processing Algorithms and their relevant GUI features
    • Support for deep learning-based object detection algorithms and conventional image processing algorithms
    • Standarized serialization format
  • Media Features::
    • Playback features
      • Support for 4K resolutions
      • Image and videos
    • Zoom/Drag features
      • Allows to validate the results of image processing algorithms
    • Data projection
      • An overlay layer that interprets the data resulting from image processing on top of the video layer
      • Support for the vizualuation of multiple objects
    • Video Shortnening
      • A feature to reduce the lenght of long videos by producing a short video containing the main parts based on the serializaiton files
      • Reduces the size of the processed datasets