Post by account_disabled on Mar 7, 2024 12:11:51 GMT
concept of the neural network above isimage recognition task. First, the pain point of image recognition is The amount of data that needs to be processed with difficulty images is too large. The picture is composed of each pixel composed of pixels. The color of the mobile phone with p, g, B. It is huge. The variable content of the picture leads to the low accuracy. If the same object is treated with the same object, and the location transformation, the objects are displayed in different attitudes. Although the object itself does not change much, it has greatly improved the picture recognition. Difficulty. And CNN can effectively solve these two problems. It can not only significantly reduce the number
of parameters and reduce the complexity; it can also be used to retain image Rich People Phone Number List features in a similar visual processing method. Even if the image is flipped and moved, it can be effectively identified. Second Visual principles to understand the principles of CNN need to understand the visual principles of human beings first. The general process of human judgment of objects can see objects (pupil intake pixels and then find the edges and directions of objects What object is the shape judgment (further abstraction. We can find that the above process is actually a low -level neural network responsible for identifying the basic characteristics of the image of the image. After the integration of multiple basic
characteristics of the image . This is the basic idea of CNN. 3. The basic principles of CNN CNN consist of three parts: convolutional layer, pooling layer, and full connection layer. Features. Pooling layer (Pooling Layer's pooling layer is responsible for significantly reduced parameter levels while retaining important feature information while reducing the computing complexity. The data processed by the pooling layer calculate the final result. Let's first take a look at the process of the local characteristics of the convolutional layer convolutional layer extraction and the extraction characteristics of human vision. Filter we call it a "convolution