In the field of deep learning among all the neural networks, CNN (Convolutional Neural Network) is the most popular and widely used.
A basic CNN model consists of two parts i.e.,
• feature extraction
CNN is a type of deep learning neural network, In a neural network, there are spatial and temporal relationships among data and like other neural networks CNN can be used to process that data.
Convolutional layers are most essential layers of CNNs. The series of convolutional layers is called layer of complexity to form a network like structure which is called a Convolutional Neural Network.
Yann LeCun developed the first CNN in 1988 when it was called LeNet. It was used for recognizing characters like ZIP codes and digits.
It is also known as ConvNets, consist of multiple layers and are mainly used for image processing and object detection.
CNN's have multiple layers that process and extract features from data.
CNNs process the data by passing it through multiple layers and extracting features to perform convolutional operations.
CNN has a convolution layer that has several filters to perform the convolution operation.
The convolutional layer consits of some activation functions like ReLU (Rectified Linear Unit) is used to perform operations on elements. The output is a rectified feature map. The rectified feature map next feeds into a pooling layer.
Pooling is a down-sampling operation that reduces the dimensions of the feature map then it converts the resulting two-dimensional arrays from the pooled feature map into a single, long, continuous, linear vector by flattening it.
Fully-connected layer: when the flattened matrix from the pooling layer is fed as an input it is called as fully-connected layer which classifies and identifies the images.
CNN is used for classification and feature-releasing into a series of trained sections. The CNN featured component typically includes integration and integration layers, while the partition component includes fully integrated and partition layers. Although CNN focuses on classifying images and accepting images as input data, for this reason, CNN is used in different fields for taking audio and video as input data.
CNN architecture typically includes six main layers named input layer, convolution layer, activation layer, pooling layer, fully-connected layer, and output layer. Also, there are some additional layers like dense layer, dropout layer, activation layer, pooling layer, etc. Input layer: It is the first layer or starting layer of the Convolutional neural network that serves as the entry point of CNN. It takes the raw pixel value of the input image including channel.
Convolutional layer: It is the second layer of CNN. All the complex mathematical convolution operations are performed in this layer to extract various features from the input images of the input layer. It also provides a feature map that contains information about the corners and edges of the images.
Activation layer: Generally, the activation layer lies between The Convolutional layer and the fully connected layers. To learn any type of relationships between the network and variables and for adding non-linearity to the network an activation function must be used. There are some popular activation functions like RELU (rectified linear unit), Softmax, TanH, and Sigmoid are commonly used in the activation layer.
Pooling layer: It serves as a bridge that connects the convolution layer and the fully-connected layer. There are three different types of pooling operations generally performed in this layer these are
Fully-connected layer: These layers are generally placed before the output layer. Input images from the previous layers are flattened before coming into this layer. It learns all the features from the series of convolution layers and pooling layers. All the mathematical operations and classification operations are performed here.
Output layer: This layer gives the result as output.
CNN consists of different types of architectures VGGNet, ResNet, DenseNet, InceptionNet, XceptionNet, MobileNet, EfficientNet etc. The transfer learning techniques play a remarkable role in computer vision. i.e., medical, surveillance, multimedia, and almost every field.
In health care systems or in medical imaging Transfer learning plays a very crucial role in standard model where the weights are train on non-medical images of natural image classification, because datasets particularly ImageNet are fine-tuned on medical imaging data.
In the field of medical imaging including MRI-Scans, CT-Scans, and X-rays transfer learning mechanisms can be implemented for training or developing a better CNN model.
Silan Software is one of the India's leading provider of offline & online training for Java, Python, AI (Machine Learning, Deep Learning), Data Science, Software Development & many more emerging Technologies.
We provide Academic Training || Industrial Training || Corporate Training || Internship || Java || Python || AI using Python || Data Science etc