c++手写数字识别knn完整

这里给出一个简单的C++ KNN数字识别的实现代码，可以对MNIST数据集进行测试：

#include <iostream>
#include <fstream>
#include <vector>
#include <cmath>
#include <cstring>

using namespace std;

const int k = 3; // KNN算法的k值
const int image_size = 784; // MNIST图片大小（28*28）
const int train_num = 60000; // 训练集大小
const int test_num = 10000; // 测试集大小

struct Data {
    vector<int> features;
    int label;
};

Data train_data[train_num];
Data test_data[test_num];

// 读取MNIST数据集
void read_mnist(string filename, Data *data, int num) {
    ifstream file(filename, ios::binary);
    if (!file) {
        cerr << "Failed to open file: " << filename << endl;
        exit(1);
    }
    unsigned char buffer[4];
    file.read((char*)buffer, 4);
    int magic_number = (buffer[0] << 24) + (buffer[1] << 16) + (buffer[2] << 8) + buffer[3];
    file.read((char*)buffer, 4);
    int n_images = (buffer[0] << 24) + (buffer[1] << 16) + (buffer[2] << 8) + buffer[3];
    for (int i = 0; i < num; i++) {
        vector<int> features(image_size);
        file.read((char*)features.data(), image_size);
        data[i].features = features;
    }
    file.close();
}

// KNN算法
int knn(Data *train_data, Data *test_data) {
    int cnt[10] = {0};
    int dist[train_num];
    for (int i = 0; i < train_num; i++) {
        int d = 0;
        for (int j = 0; j < image_size; j++) {
            int diff = train_data[i].features[j] - test_data->features[j];
            d += diff * diff;
        }
        dist[i] = d;
    }
    for (int i = 0; i < k; i++) {
        int min_dist = -1, min_index = -1;
        for (int j = 0; j < train_num; j++) {
            if (dist[j] != -1 && (min_dist == -1 || dist[j] < min_dist)) {
                min_dist = dist[j];
                min_index = j;
            }
        }
        cnt[train_data[min_index].label]++;
        dist[min_index] = -1;
    }
    int max_cnt = -1, max_index = -1;
    for (int i = 0; i < 10; i++) {
        if (cnt[i] > max_cnt) {
            max_cnt = cnt[i];
            max_index = i;
        }
    }
    return max_index;
}

int main() {
    string train_image_filename = "train-images.idx3-ubyte";
    string train_label_filename = "train-labels.idx1-ubyte";
    string test_image_filename = "t10k-images.idx3-ubyte";
    string test_label_filename = "t10k-labels.idx1-ubyte";
    read_mnist(train_image_filename, train_data, train_num);
    read_mnist(train_label_filename, train_data, train_num);
    read_mnist(test_image_filename, test_data, test_num);
    read_mnist(test_label_filename, test_data, test_num);
    int correct = 0;
    for (int i = 0; i < test_num; i++) {
        int predict_label = knn(train_data, &test_data[i]);
        if (predict_label == test_data[i].label) {
            correct++;
        }
    }
    double accuracy = (double)correct / test_num;
    cout << "Accuracy: " << accuracy << endl;
    return 0;
}

这个代码利用KNN算法对MNIST数据集进行数字识别。首先通过`read_mnist`函数读取MNIST数据集，然后利用`knn`函数对测试集中的每个样本进行预测，最后计算出准确率并输出。

需要注意的是，这个代码只是一个简单的实现，没有使用任何优化算法，因此在实际使用中识别效果可能不理想。如果想要提高识别准确率，可以使用更复杂的算法或者加入更多的特征。