Skip to content

Image segmentation and object detection

Jonathan Jones requested to merge segment-frames into master
  • New function findBlobs in rawdata.py segments images and detects objects
  • New module image.py contains low-level image processing functions
  • InferencePipeline can be configured to use only a subset (random or predetermined) of the available trial data
  • inferstate.py automatically un-mirrors early trial data
  • inferstate.py resamples all data sequences at a constant rate (30 Hz for IMU, 50 Hz for video) before processing
  • inferstate.py estimates keyframes for scene parse update given annotated state sequence
  • inferstate.py visualizes rgb & depth frame sequences along with segmentation & detected objects
  • syntax bugfix in makevideo.sh

Merge request reports