Download The MIT Artificial Intelligence Lab

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Technological singularity wikipedia , lookup

Philosophy of artificial intelligence wikipedia , lookup

AI winter wikipedia , lookup

Ethics of artificial intelligence wikipedia , lookup

Intelligence explosion wikipedia , lookup

Histogram of oriented gradients wikipedia , lookup

History of artificial intelligence wikipedia , lookup

Existential risk from artificial general intelligence wikipedia , lookup

Transcript
Visual Detection Systems
Tomaso Poggio
MIT Artificial Intelligence Laboratory — Research Directions
Object Categorization/Detection
The Problem
• Developing a general paradigm for object detection in
cluttered scenes
• Applications: target detection, visual data base search...
• Trainable system…for “any” desired object class
MIT Artificial Intelligence Laboratory — Research Directions
More on the Object Classification
System
...
...
Pedestrian
new image
Trainable
System
…..
Nonpedestrian
MIT Artificial Intelligence Laboratory — Research Directions
Learning Object Detection:
Car Detection - Training
MIT Artificial Intelligence Laboratory — Research Directions
Learning Object Detection:
Car Detection - Results
MIT Artificial Intelligence Laboratory — Research Directions
Trainable System for Object Detection:
Face Detection - Results
Training Database
1000+ Real, 3000+ VIRTUAL
50,0000+ Non-Face Pattern
Sung, Poggio 1995
MIT Artificial Intelligence Laboratory — Research Directions
Trainable System for Object Detection:
Eye Detection - Results
MIT Artificial Intelligence Laboratory — Research Directions
Trainable System for Object Detection:
Pedestrian Detection - Training
MIT Artificial Intelligence Laboratory — Research Directions
Trainable System for Object Detection:
Pedestrian Detection - Results
MIT Artificial Intelligence Laboratory — Research Directions
System Installed in Experimental
Mercedes
QuickTime™ and a
decompressor
are needed to see this picture.
A fast version, integrated
with a real-time obstacle
detection system
MPEG
MIT Artificial Intelligence Laboratory — Research Directions
QuickTime™ and a
decompressor
are needed to see this picture.
MIT Artificial Intelligence Laboratory — Research Directions
Results
The system is capable of detecting people when
they are running or walking. It is also able to detect
people when all their body parts are not detectable or
when they are slightly rotated in depth.
MIT Artificial Intelligence Laboratory — Research Directions
Results
The system is capable of detecting
partially occluded people.
MIT Artificial Intelligence Laboratory — Research Directions