Computer vision 101: Introduction

Jorge Martinez Gomez MSc.

No computer has ever been designed that is ever aware of what it is doing; but most of the time, we aren’t either – Marvin Minsky

What is computer vision?

An image is worth a thousand words – Confucius or Printers’ Ink Ad (1921)

What is computer vision?

What is computer vision?

• The ability of computers to see.
• Image/video understanding.
• Machine Vision.
• Robot Vision.
• Image/Video analysis.

​

​

What is computer vision?

Image

Sensing device

Interpreting device

Interpretations

Carpet, parade, instruments, uniforms, etc.

What is computer vision?

Image

Sensing device

Interpreting device

Interpretations

Carpet, parade, instruments, uniforms, etc.

What is computer vision?

• Computer vision is a hard problem!!

Image

Sensing device

Interpreting device

Interpretations

Carpet, parade, instruments, uniforms, etc.

What is computer vision?

• What we see

• What a computer sees

• Main goal is to bridge the gap between pixels and meaning

Why computer vision matters?

• Safety

• Health

• Suveillance

• Comfort

• Fun

• Information retrieval

Applications of computer vision

• Optical character recognition (OCR)

Applications of computer vision

• Shape and motion capture

Applications of computer vision

• 3D Reconstruction

Photosynth

http://photosynth.net

Google streetview

maps.google.com

Applications of computer vision

• 3D modeling

Agarwal, S., Snavely, N., Simon, I., Seitz, S. M., & Szeliski, R. (2009, September). Building rome in a day.

Applications of computer vision

• Object detection

Jones, M., & Viola, P. (2003). Fast multi-view face detection.

Applications of computer vision

• Object recognition

Erhan D., Szegedy C., Toshev, A., and Anguelov, D. (2014), "Scalable Object Detection using Deep Neural Networks

Applications of computer vision

• Image captioning

Karpathy, A., & Fei-Fei, L. (2015). Deep visual-semantic alignments for generating image descriptions.

Applications of computer vision

• Neural style

• Style image

• Content image

• Network output

• CNN

Gatys, L. A., Ecker, A. S., & Bethge, M. (2015). A neural algorithm of artistic style.

Code: https://github.com/jcjohnson/neural-style 

Applications of computer vision

• Object tracking

Trackers results

Ensemble method

Final ensemble best tracking result

Latest target state from ensemble

Outliers to reinitialize

Martinez J., Niebles J. Object tracking applying online ensemble of muliple trackers (2015). Master’s Thesis

Applications of computer vision

• Object tracking

Ensemble method

Final ensemble best tracking result

Applications of computer vision

• Object tracking

Ensemble method

Final ensemble best tracking result

Applications of computer vision

• People counting

Color image

Depth image

Martinez J., et at.. Automatic passenger counting for urban buses using RGBD video streams (2015).

Still a long way to go

Thank you!