Home › Claude › aifoundations › alexnet_cnn_imagenet › AlexNet's second form of data augmentation is
AlexNet's second form of data augmentation is
APCA-based intensity perturbation of the RGB channels
BCutout masking of square patches in each image
CMixup-style linear blending of pairs of images
DRandom elastic deformation of foreground objects
Answer & Solution
Correct answer: A. PCA-based intensity perturbation of the RGB channels
PCA is performed over the RGB pixel distribution; small random multiples of the principal components (weighted by their eigenvalues) are added to each image. Captures invariance to illumination intensity and colour.
Related questions
Image similarity in AlexNet's last hidden layer is measured byThe SGD update rule for AlexNet usesRemoving any single middle convolutional layer of AlexNet costs roughlyAlexNet was trained for roughly how many epochs on ImageNet?At test time dropout is approximated byAlexNet uses overlapping pooling withAlexNet's first form of data augmentation isWhy are ReLU networks faster to train than tanh networks?