Does size / resolution affect the performance of convolutional neural networks?

Question

I am creating an image classifier in which I have 66 classes and about 50,000 images, and my pc ram is 12 grams. My ram is not enough for training images, to my question, image resolution affects accuracy. I use convolutional neural networks.

Answer 1

If you use convolutional neural networks with a fully connected layer at the end to get forecast probabilities, the input level will be fixed in size, so you will need to pre-process all images of this size to feed into the network.

, , , . , .

ram, , , . OOM , , , . OOM , .