3d matrix problem index

1 visualizzazione (ultimi 30 giorni)
cristian
cristian il 3 Nov 2014
Commentato: cristian il 3 Nov 2014
really really new using matlab so here i go
if true
clear all;
cam=webcam;
img = snapshot(cam);
imwrite(img, 'a.jpg');
faceDetector = vision.CascadeObjectDetector();
videoFileReader = vision.VideoFileReader('a.jpg');
a= step(videoFileReader);
bbox= step(faceDetector, a);
m=img(bbox(2):bbox(2)+bbox(3),bbox(1):bbox(1)+bbox(3));
imshow(m);
end
the idea is to save "m" with the new bounds provided from "bbox"...thats cool i did that and "m" is the new image (only face)..the problem is "m" is a n x m unit 8 matrix image NOT like "img" a n x m x 3 unit 8 matrix image. I know about 2d (nxm) index only, i have no idea how to index a nxmx3 matrix (thats my problem i guess)... short story. Having a black and white "m" instead of the full collor "m" i wanted
any help appreciated
(sorry for my bad english)

Risposta accettata

Ced
Ced il 3 Nov 2014
Modificato: Ced il 3 Nov 2014
Hi
I'm not familiar with the output from snapshot, but concerning 3D matrices (or even higher dimensions), indexing works the same way it does for 2D matrices, just call e.g. A(1:2,2:4,1:3) instead of A(1:2,2:4). For your example, this may be something like
m = img(bbox(2):(bbox(2)+bbox(3)),bbox(1):(bbox(1)+bbox(3)),:)
since I'm guessing your 3 dimensions are color channels, and you want all three of them. If you don't specify that you want all three (through ":"), you will only get the first plane.
PS: Your english is perfectly fine.
  1 Commento
cristian
cristian il 3 Nov 2014
Thx man that completely solve my problem and i learn something new today thanks for everything

Accedi per commentare.

Più risposte (0)

Categorie

Scopri di più su Image Processing and Computer Vision in Help Center e File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by