Trying to read a csv with read_csv to load features of a CNN in a matrix

asked 2019-06-06 13:33:11 -0600

nicogc gravatar image

updated 2019-06-06 13:38:22 -0600

I am trying to train an SVM set by loading a matrix of features extracted from a CNN using CvMLData and reading the .cvs file to load it, but when I try to load.cvs the program shows an error: :

msg="C:\\opencv2412\\sources\\modules\\core\\src\\datastructs.cpp:426: error: (-211) **Storage block size is too small to fit the sequence elements in function cvSetSeqBlockSize**\n" ...

The.cvs file has the following format : 0.615685, 0.647057, 0.682354, 0.713726, 0.737255...

My.csv is 1.9 GB : (

Is there any solution for this problem?

edit retag flag offensive close merge delete

Comments

do you absolutely have to use outdated opencv 2.4 for this ?

api has changed significantly.

berak gravatar imageberak ( 2019-06-06 13:37:55 -0600 )edit

Yes it is necessary. Unfortunately, I can not update the version because I am doing my final year project with the code that the university has given me, which implements a training and classification system, I have to adapt it and improve it. That's why it's very difficult for me to change all the code so that it can work with a new version of opencv

nicogc gravatar imagenicogc ( 2019-06-06 13:50:12 -0600 )edit