Please, advice how to pre-processing the image labelling.

I'm initially trying to use alexnet to train the image in the kaggle competition.

there are 4 defect classes(1 ~ 4) .

Problem is that the images which don't have defect were not labeled, so these are not included in the csv file.

How can I delete the unlabeled images and make the label on neural network?

This is the website :

https://www.kaggle.com/c/severstal-steel-defect-detection/data?select=train.csv

steel_ds = imageDatastore('*.jpg');

information = readtable('train.csv');

Answer 1

[yanqi liu]

loop for table

use exist(filename, 'file') to judge the file is or not exist