Visual Intelligence & Video Analytics
Download as PPTX, PDF2 likes915 views
This document discusses modern visual information workflows and key technologies driving changes in visual data capture and analysis. It outlines opportunities in transportation monitoring, human behavior classification, visual search, and health diagnostics using computer vision and artificial intelligence. While computational power, data storage, and input sources are increasing, technical challenges remain around privacy, reliability, and obtaining sufficient data. Overall, proliferation of cameras and imaging devices is leading to more visual data that can be analyzed to gain insights.
Visual Intelligence & Video Analytics
- 1. Visual Intelligence & Video Analytics
- 2. Modern Visual Information Work Flow Physical World Decision
- 3. • Manual capture at ground level • Limited to where he can travel A • Auto capture at elevated level • Limited to selected locations B Key Milestones For Capturing Visual Information • Auto capture from top • Freedom to view any part of Earth D • Auto capture from top • Freedom to roam with limited range C
- 4. A B DC Key Technologies Which Drives Changes • Invention of CCD/CMOS image sensor • Broadband communication • Improved SW & Avionics • Solid-state IMU & Gyroscope • Composite material • Fuel cell battery • High efficiency solar panel • Improved electronics • Chemical-based photo-sensitive film • Miniaturized opticsTechnological changes have bring forth new Capabilities
- 5. Google GlassSmart Watch Wearable Camera Pill Camera In-Vehicle Camera Mobile Phone Camera Proliferation of imaging devices are leading to an ever increasing amount of visual information What’s Next ?
- 6. Mckinsey & Company, 2013 May, 12 Disruptive technologies: Advances that will transform the global economy
- 7. Opportunity : Number 5, Transportation & Traffic LPR Speeding Bus lane intrusion Smoky exhaust Traffic condition
- 8. Opportunity : Number 4, Human Behavior Classification Fight Aggressive behavior Anger ShopliftingCrowd formation
- 9. Opportunity : Number 3, Visual Search (Rigid/Semi-Rigid Obj)
- 10. Opportunity : Number 2, Facial Recognition
- 11. Opportunity : Number 2, Facial Recognition
- 12. Opportunity: Number 1, Health Diagnostic
- 13. Computation Power Data Storage Input Sources Bandwidth Algorithm • ML, AI, DM • Tradeoffs Challenges (Technical)
- 14. Challenges (Operational) Privacy • Problem getting sufficient data Reliability • Is it good enough ?
- 15. For many species, survival depends on whether it could see it’s prey and predator
Editor's Notes
- #6: Bring forth by storage and communication improvements.
- #9: Human Behavior analysis Aggressive behavior, Burger Intrusion into secure area Urbanization, Loss of wealth
- #10: Object & Behavior Search Able to Identify a location from background image
- #12: Facial recognition Who is this person ? Where else can I find him ? Info about his past Cloud based analytics