matlab基于CNN卷积神经网络猫狗猪动物识别系统.zip代码102

clc; clear all; close all; warning off all; % 获取字符集 fd = fullfile(pwd, 'images', 'dbx'); imds = imageDatastore(fd, ... 'IncludeSubfolders',true,'LabelSource','foldernames'); classes = categories(imds.Labels); [imdsTrain,imdsValidation] = splitEachLabel(imds,rateTrainFiles,'randomize'); % 输入输出设置 image_size = [117 117 1]; class_number = length(classes); % 数据维数对应 inputSize = layers(1).InputSize; imdsTrain = augmentedImageDatastore(inputSize(1:2),imdsTrain,'ColorPreprocessing','rgb2gray'); augimdsValidation = augmentedImageDatastore(inputSize(1:2),imdsValidation,'ColorPreprocessing','rgb2gray'); % 设置训练参数 MaxEpochs = 10; InitialLearnRate = 0.001; MiniBatchSize = 50; % 设置训练环境 ExecutionEnvironment = 'auto'; options_train = trainingOptions('sgdm',... 'MaxEpochs',MaxEpochs,... 'Verbose',true,'MiniBatchSize', MiniBatchSize,... 'Plots','training-progress',... 'ValidationData',augimdsValidation , ... 'ValidationFrequency',10, ... 'ExecutionEnvironment', ExecutionEnvironment); net_file = 'cnn_net.mat'; if exist(net_file, 'file') load(net_file); else % 训练保存 net = trainNetwork(imdsTrain, layers, options_train); % 存储网络 save(net_file, 'net'); end % 测试 while 1 filePath = OpenImageFile(); if isequal(filePath, 0) break; end im = imread(filePath); x = im; % 维度对应 inputSize = layers(1).InputSize; if ~isequal(size(x), inputSize(1:2)) x = imresize(x, inputSize(1:2), 'bilinear'); end xw = zeros(inputSize(1),inputSize(2), 1, 1); xw(:,:,1,1) = x(:,:,1); t1 = cputime; % 评测 yw = classify(net,xw); t2 = cputime; res = sprintf('\s\n识别结果为%s', t2-t1, char(yw)); figure(1); clf; imshow(im); title(res, 'FontSize', 16); pause(1); end %% 欢迎沟通交流Q：3752243968 %% 欢迎沟通交流Q：3752243968 %% 欢迎沟通交流Q：3752243968