From the course: Deep Learning: Image Recognition
Join today to access over 24,800 courses taught by industry experts.
Convolutional neural networks (CNNs) - Python Tutorial
From the course: Deep Learning: Image Recognition
Convolutional neural networks (CNNs)
- Now we have prepared our data, let's go ahead and talk about convolutional neural networks, in other words, CNNs. So CNNs are a class of deep neural networks primarily used for image classification and recognition. They automatically and adaptively learn spatial hierarchies of features for from the input images that we provide. Well, where is this use? This is widely used in computer vision tasks such as imagery cognition, object detection, and segmentation. While the CNN architecture consists of different layers, it starts with an input layer so accepts input images, for example, 32x32 RGB images for our dataset, in this case, CIFAR-10, and then it has convolutional layers. It applies convolutional filters to extract features from the input images. It has an activation function, ReLU. It introduces non-linearity to our model. Next, we have pooling layers where it downssamples the feature maps to reduce spatial dimensions and therefore it eases the computational complexity which…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.