Application Areas of CNNs

Now that we understand the architecture of CNNs, let's look at some applications. In general, CNNs are great for data that has a spatial structure. Examples of types of data that has a spatial structure are sound, images, video, and text.

In natural language processing, CNNs are used for various tasks such as sentence classification. One example is the task of sentiment classification, where a sentence is classified as belonging to a predetermined group of classes.

As discussed earlier, CNNs are applied at the character level to classification tasks such as sentiment classification, especially on noisy datasets such as social media posts.

CNNs are more commonly applied in computer vision. Here are some applications in this area:

• Facial recognition

  Most social networking sites employ CNNs to detect faces and subsequently perform tasks such as tagging.

Figure 4.22: Facial recognition

• Object detection

  Similarly, CNNs are able to detect...