Python Deep Learning - Second Edition

Deep learning and sensors

Now that we have an idea what sensors the vehicle uses, let's see how to apply deep learning to the raw sensor data. First, we'll do this for the camera. In Chapter 5, Advanced Computer Vision, we discussed how to use CNNs in two advanced vision tasks: object detection and semantic segmentation. To recap, object detection creates a bounding box around different classes of objects detected in the image. Semantic segmentation assigns a class label to every pixel of the image. We can use segmentation to detect the exact shape of the road surface and the lane markings on the camera image. We can use object detection to classify and localize the objects of interest in the environment. These include other vehicles, pedestrians, ...