questions about "surf.surf" function

[Antonis Polykratis]

Hello,

I have two questions

1) I try to understand the following line of code that I saw in a text:
surf.surf(image)[:, 5:]
Is there any known (general) reason for the above slicing? Why not
surf.surf(image)?

2) surf.surf(image) is returning empty result in some photos. I think that
these photos are small.
I would expect that in these cases the features would be less than normal
but not zero. (So in these
cases I have to add logic to exclude those images).
Any comment on that?

Thanks for your time.
Antonis

[Luis Pedro Coelho]

Hi Antonis,

Post by Antonis Polykratis
surf.surf(image)[:, 5:]
Is there any known (general) reason for the above slicing? Why not
surf.surf(image)?

The first few elements contain the position (y,x), the scale, the score,
the Laplacian. These are ignored by the code above.

See:
http://mahotas.readthedocs.org/en/latest/surf.html

Post by Antonis Polykratis
2) surf.surf(image) is returning empty result in some photos. I think that
these photos are small.
I would expect that in these cases the features would be less than normal
but not zero. (So in these
cases I have to add logic to exclude those images).
Any comment on that?

This probably means that no interest points were detected. Perhaps, try
dense sampling:

see
http://mahotas.readthedocs.org/en/latest/surf.html?highlight=dense#api-documentation

HTH
Luis

[Antonis Polykratis]

Thank you Luis!

A.


΀η Σάββατο, 7 Μαρτίου 2015 - 5:56:54 ÎŒ.ÎŒ. UTC+2, ο χρήστης Antonis

Post by Antonis Polykratis
Hello,
I have two questions
surf.surf(image)[:, 5:]
Is there any known (general) reason for the above slicing? Why not
surf.surf(image)?
2) surf.surf(image) is returning empty result in some photos. I think that
these photos are small.
I would expect that in these cases the features would be less than normal
but not zero. (So in these
cases I have to add logic to exclude those images).
Any comment on that?
Thanks for your time.
Antonis

[Antonis Polykratis]

Thank you Luis!

A.