feat(knowing app): stm32f4 svc iris demo

APP_Framework/Applications/knowing_app/Kconfig

@@ -3,4 +3,5 @@ menu "knowing app"
     source "$APP_DIR/Applications/knowing_app/face_detect/Kconfig" 
     source "$APP_DIR/Applications/knowing_app/instrusion_detect/Kconfig" 
     source "$APP_DIR/Applications/knowing_app/helmet_detect/Kconfig" 
+    source "$APP_DIR/Applications/knowing_app/iris_ml_demo/Kconfig" 
 endmenu

APP_Framework/Applications/knowing_app/iris_ml_demo/Kconfig

@@ -0,0 +1,3 @@
+config IRIS_ML_DEMO
+    bool "enable apps/iris ml demo"
+    default n

APP_Framework/Applications/knowing_app/iris_ml_demo/SConscript

@@ -0,0 +1,9 @@
+from building import *
+
+cwd     = GetCurrentDir()
+src     = Glob('*.c') + Glob('*.cpp')
+CPPPATH = [cwd]
+
+group = DefineGroup('Applications', src, depend = ['IRIS_ML_DEMO'], LOCAL_CPPPATH = CPPPATH)
+
+Return('group')

APP_Framework/Applications/knowing_app/iris_ml_demo/SVCModel.h

@@ -0,0 +1,78 @@
+/**
+ * SVC model trained by iris dataset
+ */
+#include <stdarg.h>
+
+/**
+ * Compute kernel between feature vector and support vector.
+ * Kernel type: linear
+ */
+float compute_kernel(float *x, ...)
+{
+    va_list w;
+    va_start(w, 4);
+    float kernel = 0.0;
+
+    for (int i = 0; i < 4; i++) {
+        kernel += x[i] * va_arg(w, double);
+    }
+
+    return kernel;
+}
+
+/**
+ * Predict class for features vector
+ */
+int predict(float *x)
+{
+    float kernels[27] = {0};
+    float decisions[3] = {0};
+    int votes[3] = {0};
+    kernels[0] = compute_kernel(x, 5.1, 3.3, 1.7, 0.5);
+    kernels[1] = compute_kernel(x, 4.8, 3.4, 1.9, 0.2);
+    kernels[2] = compute_kernel(x, 4.5, 2.3, 1.3, 0.3);
+    kernels[3] = compute_kernel(x, 6.9, 3.1, 4.9, 1.5);
+    kernels[4] = compute_kernel(x, 6.3, 3.3, 4.7, 1.6);
+    kernels[5] = compute_kernel(x, 6.1, 2.9, 4.7, 1.4);
+    kernels[6] = compute_kernel(x, 5.6, 3.0, 4.5, 1.5);
+    kernels[7] = compute_kernel(x, 6.2, 2.2, 4.5, 1.5);
+    kernels[8] = compute_kernel(x, 5.9, 3.2, 4.8, 1.8);
+    kernels[9] = compute_kernel(x, 6.3, 2.5, 4.9, 1.5);
+    kernels[10] = compute_kernel(x, 6.8, 2.8, 4.8, 1.4);
+    kernels[11] = compute_kernel(x, 6.7, 3.0, 5.0, 1.7);
+    kernels[12] = compute_kernel(x, 6.0, 2.7, 5.1, 1.6);
+    kernels[13] = compute_kernel(x, 5.4, 3.0, 4.5, 1.5);
+    kernels[14] = compute_kernel(x, 5.1, 2.5, 3.0, 1.1);
+    kernels[15] = compute_kernel(x, 4.9, 2.5, 4.5, 1.7);
+    kernels[16] = compute_kernel(x, 6.5, 3.2, 5.1, 2.0);
+    kernels[17] = compute_kernel(x, 6.0, 2.2, 5.0, 1.5);
+    kernels[18] = compute_kernel(x, 6.3, 2.7, 4.9, 1.8);
+    kernels[19] = compute_kernel(x, 6.2, 2.8, 4.8, 1.8);
+    kernels[20] = compute_kernel(x, 6.1, 3.0, 4.9, 1.8);
+    kernels[21] = compute_kernel(x, 7.2, 3.0, 5.8, 1.6);
+    kernels[22] = compute_kernel(x, 6.3, 2.8, 5.1, 1.5);
+    kernels[23] = compute_kernel(x, 6.0, 3.0, 4.8, 1.8);
+    kernels[24] = compute_kernel(x, 6.3, 2.5, 5.0, 1.9);
+    kernels[25] = compute_kernel(x, 6.5, 3.0, 5.2, 2.0);
+    kernels[26] = compute_kernel(x, 5.9, 3.0, 5.1, 1.8);
+    decisions[0] = 1.452844496978 + kernels[0] * 0.67075289031 + kernels[2] * 0.077097563476 + kernels[14] * -0.747850453786;
+    decisions[1] = 1.507713125178 + kernels[0] * 0.043820415076 + kernels[1] * 0.159872086718 + kernels[15] * -0.203692501794;
+    decisions[2] = 6.78097118511 + kernels[3] + kernels[4] + kernels[5] + kernels[6] + kernels[7] + kernels[8] + kernels[9] +
+                   kernels[10] * 0.243261886421 + kernels[11] + kernels[12] + kernels[13] - kernels[15] - kernels[16] -
+                   kernels[17] - kernels[18] - kernels[19] - kernels[20] + kernels[21] * -0.437859817863 - kernels[22] -
+                   kernels[23] + kernels[24] * -0.645105347981 + kernels[25] * -0.160296720576 - kernels[26];
+    votes[decisions[0] > 0 ? 0 : 1] += 1;
+    votes[decisions[1] > 0 ? 0 : 2] += 1;
+    votes[decisions[2] > 0 ? 1 : 2] += 1;
+    int val = votes[0];
+    int idx = 0;
+
+    for (int i = 1; i < 3; i++) {
+        if (votes[i] > val) {
+            val = votes[i];
+            idx = i;
+        }
+    }
+
+    return idx;
+}

APP_Framework/Applications/knowing_app/iris_ml_demo/iris.csv

@@ -0,0 +1,7 @@
+5.1,3.5,1.4,0.2
+6.4,3.2,4.5,1.5
+5.8,2.7,5.1,1.9
+7.7,3.8,6.7,2.2
+5.5,2.6,4.4,1.2
+5.1,3.8,1.9,0.4
+5.8,2.7,3.9,1.2

APP_Framework/Applications/knowing_app/iris_ml_demo/iris_ml_demo.c

@@ -0,0 +1,59 @@
+#include <string.h>
+#include <transform.h>
+
+#include "SVCModel.h"
+
+#define FEATURE_NUM 4
+#define CSV_PATH "/csv/iris.csv"
+#define CSV_BUFFER_SIZE (1 * 1024)
+
+static float data[10][FEATURE_NUM] = {};
+static int data_len = 0;
+
+void simple_CSV_read()
+{
+    int fin;
+    int col = 0;
+    char buffer[CSV_BUFFER_SIZE] = "";
+    char *tmp = "";
+    char *delim = ",\n ";
+
+    fin = open(CSV_PATH, O_RDONLY);
+    if (!fin) {
+        printf("Error open file %s", CSV_PATH);
+        exit(-1);
+    }
+    read(fin, buffer, sizeof(buffer));
+    close(fin);
+
+    data_len = 0;
+    for (tmp = strtok(buffer, delim); tmp && *tmp; col++, tmp = strtok(NULL, delim)) {
+        if (0 == col % FEATURE_NUM) {
+            // printf("\n");
+            data_len++;
+            col = 0;
+        }
+        data[data_len - 1][col] = atof(tmp);
+        // printf("%.4f ", data[data_len - 1][col]);
+    }
+    // printf("\n");
+}
+
+void iris_SVC_predict()
+{
+    int result;
+
+    simple_CSV_read();
+
+    for (int i = 0; i < data_len; i++) {
+        result = predict(data[i]);
+        printf("data %d: ", i + 1);
+        for (int j = 0; j < FEATURE_NUM; j++) {
+            printf("%.4f ", data[i][j]);
+        }
+        printf("result: %d\n", result);
+    }
+}
+#ifdef __RT_THREAD_H__
+MSH_CMD_EXPORT(iris_SVC_predict, iris predict by SVC);
+#endif