Fig. 5: The typical CNN architecture for image-classification tasks. | Light: Science & Applications

Fig. 5: The typical CNN architecture for image-classification tasks.

From: Deep learning in optical metrology: a review

Fig. 5

a The typical CNN architecture for image classification tasks consists of the input layer, convolutional layers, fully connected layers, and output prediction. b Convolution operation. c Pooling operation

Back to article page