What's wrong with my feature extraction code?
Información
La pregunta está cerrada. Vuélvala a abrir para editarla o responderla.
Mostrar comentarios más antiguos
I am getting same descriptors for different images.
I tested it by:
image1 = imread('C:\Users\Explorer\Documents\MATLAB\ASL_signs\A.jpg');
image2 = imread('C:\Users\Explorer\Documents\MATLAB\ASL_signs\B.jpg');
v1=Descrip(image1);
v2=Descrip(image2);
SUM=sum(v1-v2)
_____________________________________________________________________________________
Output :
SUM =
0
1 comentario
Explorer
el 12 de Feb. de 2014
Respuestas (2)
David Young
el 12 de Feb. de 2014
0 votos
Descrip writes then reads contour.jpg, and then computes the result from that. One way to set about debugging would be to see whether contour.jpg is different for the two images. That will help pin down where the problem lies.
4 comentarios
Explorer
el 12 de Feb. de 2014
David Young
el 13 de Feb. de 2014
The idea was that you run Descrip with one input image, look at contour.jpg, run it again with the other input image, and look at contour.jpg again. Are they different?
Image Analyst
el 12 de Feb. de 2014
Editada: Image Analyst
el 12 de Feb. de 2014
0 votos
Don't use image as a variable name - it's the name of a built-in function.
5 comentarios
Explorer
el 13 de Feb. de 2014
Image Analyst
el 13 de Feb. de 2014
Editada: Image Analyst
el 13 de Feb. de 2014
What are you using now: Descrip or Descrip2? Doesn't stepping through it with the debugger reveal where the problem is????
Image Analyst
el 13 de Feb. de 2014
If image A equals image B then I think they should be the same. You didn't attach either image so only you can know that. What did stepping through with the debugger reveal to you?
Walter Roberson
el 13 de Feb. de 2014
Watch out for the case where v1 and v2 are being returned as unsigned integers; in that case if v2 > v1 you would get "clipping" at 0 instead of getting a negative number.
La pregunta está cerrada.
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!