Image-Classification-with-localization

• In this project i did Classification with localization (aka object localization)
• I used portion of Caltech-101 dataset this protion was ['airplanes', 'Faces', 'Motorbikes']
• Version 2:
  • functions that I used to parse data and show image all in my_helper_funcs.py file
  • I trained two transfer learning models:
    • first one was with NasNet as base model
      • training class accuracy: 100%
      • training bounding box loss: 1551.15
      • Validation accuracy: 100%
      • training bounding box loss: 1293.05
    • second model was with VGG-16 as base model
      • training class accuracy: 0.9984%
      • training bounding box loss: 688.58
      • Validation accuracy: 100%
      • training bounding box loss: 802.32
• Version 3:
  • https://drive.google.com/drive/folders/1I55CnyZbGj2Gsg5IVqeIaKitRl8WpA97?usp=drive_link

- Note: I did not do image Augmentation yet