//识别
private void label3_Click(object sender, EventArgs e)
{
#region 准备字典
InitDict();
var dict = Dict.ToArray();
#endregion
#region 把每个笔画录入矩阵数据
var pointdd = new Point[1][];
var traindata = new MatrixData();
for (int i = 0; i < ArrayPtBuff.Count; i++)
{
pointdd[0] = ArrayPtBuff[i].ToArray();
traindata.RecordOne(pointdd);
}
#endregion
//识别的数据
var test = GetDim1Matrix(traindata);
//笔画识别结果
var sortresult = "";
#region 让神经网络加载训练结果并进行识别
var hbpn = new HBPN(File.ReadAllText("GData.json"));
for (int i = 0; i < test.GetLength(0); i++)
{
var sim0 = hbpn.sim(test.GetColumn(i));
int like = 0;
var distant = double.MaxValue;
for (int k = 0; k < sim0.Length; k++)
{
if (Math.Abs(sim0[k] - 1) > distant)
{
continue;
}
distant = Math.Abs(sim0[k] - 1);
like = k;
}
//result += "㇐㇑㇀㇏㇕㇗㇡"[like];
sortresult += "123455555555555"[like];
}
#endregion
#region 这里计算字典中每个字与识别结果的相似度
for (int i = 0; i < dict.Length; i++)
{
dict[i].distance = EditorDistance.Levenshtein(dict[i].Code[0], sortresult);
}
#endregion
#region 根据相似度,对字典进行排序
for (int i = 0; i < dict.Length; i++)
{
for (int j = i + 1; j < dict.Length; j++)
{
if (dict[i].distance < dict[j].distance)
{
var temp = dict[i];
dict[i] = dict[j];
dict[j] = temp;
}
}
}
#endregion
//字符识别结果
var result = "";
#region 根据相似度排序
for (int i = 0; i < 20; i++)
{
result += dict[i].TheChar.ToString();
}
#endregion
MessageBox.Show(result, sortresult);
Clear();
traindata.Clear();
}
private void button1_Click(object sender, EventArgs e)
{
MatrixData.Clear();
MessageBox.Show("OK");
return;
}
