Video alignment for change detection



This  web page contains additional material of the manuscript
Video Alignment for change detecion by F. Diego, D. Ponsa, J. Serrat and A. López.

Sequences

  1. campus1

  2. highway1

  3. backroad1

  4. street

  5. night

    1. campus2

    2. highway2

    3. backroad2



Sequences

Normal driving conditions

1. CAMPUS1

Original videos

Synchronization

Video Alignment

Difference
Fusion

2. HIGHWAY1

Original videos

Synchronization

Video Alignment

Difference
Fusion

3. BACKROAD1

Original videos

Synchronization

Video Alignment

Difference
Fusion


4. STREET

Original videos

Synchronization

Video Alignment

Difference
Fusion

5. NIGHT

Original videos

Synchronization

Video Alignment

Difference
Fusion


Frequent and sharp acceleration and braking

6. CAMPUS2
Original videos

 
Synchronization

Video Alignment

Difference
 Fusion

7. HIGHWAY2

Original videos

Synchronization

Video Alignment

Difference
Fusion

8. BACKROAD2

Original videos

Synchronization

Video Alignment

Difference
Fusion

Vehicle detection


 vehicleDectection
                                                  Green box: ground truth
                                                  Yellow box: detection

 Outdoor Surveillance

vehicleDectection

Comparison w.r.t. Sand & Teller algorithm [1]
[1] Peter Sand and Seth Teller, Video Matching in ACM Transactions on Graphics (Proc. SIGGRAPH), Vol. 22, No. 3, July 2004, pp. 592-599.


Original videos

This text will be replaced
Sand & Teller algorithm

This text will be replaced
Our method

This text will be replaced




Synthetic road

The synthetic results have been used to validate the temporal and spatial registration with regard to a ground truth.


Original videos


Synchronization


Video Alignment




Pitch, yaw and roll angles estimation versus ground truth. The maximum difference is less than 0.2 degrees.

 

Estimated temporal correspondence versus ground truth.




Last update: September 14th, 2010


Made with Nvu