卷积神经网络数字图片识别

哔哩哔哩教程：【卷积神经网络数字图片识别（MATLAB）】https://www.bilibili.com/video/BV1n1421y7kZ?vd_source=471188e4bb1a707946690b75051d386c

一、搭建卷积神经网络模型（数字图片识别模型）

% 步骤1：准备数据
digitDatasetPath = fullfile(matlabroot,'toolbox','nnet','nndemos', ...
    'nndatasets','DigitDataset');
digitData = imageDatastore(digitDatasetPath, ...
    'IncludeSubfolders',true,'LabelSource','foldernames');

% 数据增强
augmenter = imageDataAugmenter( ...
    'RandRotation',[-10,10], ...  % 随机旋转角度范围为 -10 到 10 度
    'RandXTranslation',[-3,3], ... % 随机水平平移范围为 -3 到 3 像素
    'RandYTranslation',[-3,3]);    % 随机垂直平移范围为 -3 到 3 像素

digitDataAugmented = augmentedImageDatastore([28 28],digitData, ...
    'DataAugmentation',augmenter);

% 划分数据集
[trainImgs,testImgs] = splitEachLabel(digitData,0.7,'randomized');

% 步骤2：定义网络结构
layers = [
    imageInputLayer([28 28 1])
    
    convolution2dLayer(3,16,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(2,'Stride',2)
    
    convolution2dLayer(3,32,'Padding','same')
    batchNormalizationLayer
    reluLayer
    
    maxPooling2dLayer(2,'Stride',2)
    
    fullyConnectedLayer(10)
    softmaxLayer
    classificationLayer
];

% 步骤3：训练网络
options = trainingOptions('sgdm', ...
    'MaxEpochs',5, ...
    'MiniBatchSize',64, ...
    'InitialLearnRate',0.01, ...
    'Shuffle','every-epoch', ...
    'Verbose',false, ...
    'Plots','training-progress');

net = trainNetwork(trainImgs,layers,options);

% 步骤4：测试网络
predictedLabels = classify(net,testImgs);
accuracy = sum(predictedLabels == testImgs.Labels) / numel(testImgs.Labels);
fprintf('测试准确率: %.2f%%\n', accuracy * 100);

% 保存训练好的网络
save('trainedDigitNet.mat', 'net');

二、调用模型函数代码编写

% 加载训练好的网络
load('trainedDigitNet.mat', 'net');
% 读取输入的数字图片
inputImage = imread('C:\Users\19350\Desktop\920.jpg'); % 替换为你的图片路径
inputImage = rgb2gray(inputImage); % 将彩色图像转换为灰度图像
inputImage = imresize(inputImage, [28, 28]); % 将图像大小调整为28x28像素

% 使用模型进行预测
predictedLabel = classify(net, inputImage);

% 显示预测结果
disp(['识别结果为: ', char(predictedLabel)]);

三、创建APP可视化卷积神经网络数字图片识别界面

classdef DigitRecognizerApp < matlab.apps.AppBase

    % Properties that correspond to app components
    properties (Access = public)
        UIFigure                matlab.ui.Figure
        LoadImageButton         matlab.ui.control.Button
        UIAxes                  matlab.ui.control.UIAxes
        RecognitionResultLabel  matlab.ui.control.Label
    end

    properties (Access = private)
        net % Pre-trained network
    end

    methods (Access = private)

        % Button pushed function: LoadImageButton
        function LoadImageButtonPushed(app, event)
            [file, path] = uigetfile({'*.jpg;*.png;*.bmp', 'Image Files'});
            if isequal(file, 0)
                return;
            end
            imagePath = fullfile(path, file);
            inputImage = imread(imagePath);
            inputImage = rgb2gray(inputImage);
            inputImage = imresize(inputImage, [28, 28]);

            % Display the image
            imshow(inputImage, 'Parent', app.UIAxes);

            % Use the trained network to classify the image
            predictedLabel = classify(app.net, inputImage);

            % Display the result
            app.RecognitionResultLabel.Text = ['识别结果为: ', char(predictedLabel)];
        end
    end

    methods (Access = public)

        % Construct app
        function app = DigitRecognizerApp

            % Create and configure components
            createComponents(app)

            % Load the pre-trained network
            app.net = load('trainedDigitNet.mat', 'net');
            app.net = app.net.net;
        end

        % Code that executes before app deletion
        function delete(app)
            % Delete UIFigure when app is deleted
            delete(app.UIFigure)
        end
    end

    methods (Access = private)

        % Create and configure components
        function createComponents(app)

            % Create UIFigure
            app.UIFigure = uifigure;
            app.UIFigure.Position = [100 100 400 300];
            app.UIFigure.Name = '数字识别应用';

            % Create LoadImageButton
            app.LoadImageButton = uibutton(app.UIFigure, 'push');
            app.LoadImageButton.ButtonPushedFcn = createCallbackFcn(app, @LoadImageButtonPushed, true);
            app.LoadImageButton.Position = [150 220 100 30];
            app.LoadImageButton.Text = '加载图像';

            % Create UIAxes
            app.UIAxes = uiaxes(app.UIFigure);
            app.UIAxes.Position = [50 50 150 150];

            % Create RecognitionResultLabel
            app.RecognitionResultLabel = uilabel(app.UIFigure);
            app.RecognitionResultLabel.Position = [220 50 150 30];
            app.RecognitionResultLabel.Text = '识别结果为: ';
        end
    end
end

四、最后结果显示