Discussion

Both GoogleNet and AlexNet image recognition system can identify an object, but they cannot identify multiple objects.

Group Discussion Points

• How can image recognition system be improved so that it can detect multiple objects? What kinds of methods do you think the system would need? There might be:
  • Object localization System
  • Segmentation System
  • Augmentation System
  • Object Tracking
  • etc.,
• Why would we need to detect multiple objects within a screen? Where would this system be implemented?