Page 270 - AI Ver 1.0 Class 10
P. 270

Dog Dog


                                                                                                      Bird Bird

                                                                                                      Cat Cat


                    Pixels of image fed as input  Input Layer       Hidden Layer         Output Layer

                             Input Image               Image is processed                      Output probability
                           (Blue Jay)                 using Convolutional                        Beagel - 10%
                                                        Neural Network                          Blue Jay - 70 %
                                                                    Maine Coon - 20%



                               Brainy Fact



                    Convolutional neural networks (ConvNets) were first introduced in the 1980s by Yann LeCun, a computer science
                    researcher. Its early version called LeNet (after LeCun), were used to recognize handwritten digits. It found its
                    use in postal services to read zip codes on envelopes and in banking/financial sectors to read digits on cheques.





              Layers of Convolutional Neural Network
              Convolutional Neural Network is made up of multiple layers of artificial neurons. It uses mathematical functions
              to calculate the weighted sum of multiple inputs and generate the desired outputs. Let us discuss in detail about

              the different layers present in Convolutional Neural Network.
              Convolution Layer
              Convolution is the first layer of a CNN and is also known as Feature Extractor Layer. The main purpose of this
              layer is to extract the high-level features from the input image to perform operations such as edge detection,

              blur and sharpen by applying filters.
              Depending on the type of AI model, a Convolution Layer may be made up of two or more layers. In that case
              the first Convolution Layer is responsible for capturing the Low-Level features such as edges, colour, gradient
              orientation, etc. With added layers, the architecture adapts to the High-Level features and a whole network of

              understanding of images in the dataset.
              This layer deals with the Convolution process of handling an image with several types of kernels to provide
              different features to the whole system. Each convolutional kernel is used to generate a feature map based on
              input provided. Feature map is also known as the activation map. Feature maps have multiple uses like:

                 •  The output of the filter applied to the previous layer is trapped by the feature map.
                 •  It helps in reducing the size of the image so that it can be processed easily.

                 •    It helps in focusing on the important features of the images like eyes, nose etc. so that it can be processed
                  efficiently.



                        268   Touchpad Artificial Intelligence-X
   265   266   267   268   269   270   271   272   273   274   275