Ask Your Question
0

Compare functions after normalize

asked 2017-04-05 00:20:43 -0500

Hey, guys! I'm doing a video reading and for each frame I need to calculate the average standard deviation and the variance. (I did these calculations without problems using the openCv method).

Now, after normalizing each of the frames, I need to compare the two functions (the original and normalized) using metric L1 (manhattan distance).

Does anyone have any ideas or suggestions on how this can be done?

edit retag flag offensive close merge delete

Comments

just curious, how did you do the normalization, and what are you trying to achieve, in general ?

berak gravatar imageberak ( 2017-04-05 01:21:55 -0500 )edit

I used the function http://docs.opencv.org/2.4.8/modules/core/doc/operations_on_arrays.html?highlight=normalize#cv2.normalize (Core.Normalize()).

I could not do question four.;

(DataMeasures on Image Sequences) Define three different data measures Di (t), i = 1, 2, 3, for analysing image sequences. Your program should do the following: 1. Read as input an image sequence (e.g. in VGA format) of at least 50 frames. 2. Calculate your data measures Di (t), i = 1, 2, 3, for those frames. 3. Normalize the obtained functions such that all have the same mean and the same variance. 4. Compare the normalized functions by using the L1-metric. Discuss the degree of structural similarity between your measures in dependence of the chosen input sequence of images.

marceloopencv gravatar imagemarceloopencv ( 2017-04-05 14:14:07 -0500 )edit

1 answer

Sort by ยป oldest newest most voted
1

answered 2017-04-05 01:24:41 -0500

berak gravatar image

the comparison can be done with a simple norm()

like:

double distance = norm( a, b, Core.NORM_L1 );
edit flag offensive delete link more

Comments

Thank you so much!

marceloopencv gravatar imagemarceloopencv ( 2017-04-05 20:24:43 -0500 )edit
Login/Signup to Answer

Question Tools

1 follower

Stats

Asked: 2017-04-05 00:20:43 -0500

Seen: 84 times

Last updated: Apr 05 '17