我可以從 Azure 認知服務中檢索已識別的置信度嗎？

Question

我正在以 C 尖銳的 window 形式對文本應用程序進行語音。 它工作正常並在視覺工作室中運行，但是

我正在使用此代碼來識別使用 Microsoft Azure 認知服務。 一旦整個事情都被識別出來，我可以在 c 尖銳 window 表格中獲得信心分數嗎？

我該如何解決這個問題？

我的代碼：

using System;
using System.Collections.Generic;
using System.ComponentModel;
using System.Data;
using System.Drawing;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using System.Windows.Forms;
using Microsoft.CognitiveServices.Speech;
using Microsoft.CognitiveServices.Speech.Audio;
using System.IO;
using System.Threading;

namespace WindowsFormsApp2
{
    public partial class Form1 : Form
    {
        private bool isRecognizing = false;
        private SpeechRecognizer recognizer;


        public Form1()
        {
            InitializeComponent();
        }

        private void initRecognizer()
        {
            
            SpeechConfig config = SpeechConfig.FromSubscription("key", "region");
            if (Properties.Settings.Default.Punctuation)
            {
                config.SetServiceProperty("punctuation", "explicit", ServicePropertyChannel.UriQueryParameter);
            }
            //AudioConfig audioConfig = AudioConfig.FromMicrophoneInput();
            recognizer = new SpeechRecognizer(config/*, audioConfig*/);
            recognizer.Recognized += SpeechRecognizer_Recognized;
        }
        private void Form1_Load(object sender, EventArgs e)
        {
            initRecognizer();

        }

        private void SpeechRecognizer_Recognized(object sender, SpeechRecognitionEventArgs e)
        {
            if (e.Result.Reason == ResultReason.RecognizedSpeech)
            {
                if (e.Result.Text.ToLower().Equals("new line") || e.Result.Text.ToLower().Equals("newline"))
                {
                    SendKeys.SendWait(Environment.NewLine);
                }
                else
                {
                    SendKeys.SendWait(e.Result.Text);
                }

            }
        }


        private void Startstop()
        {
            if (isRecognizing)
            {
                recognizer.StopContinuousRecognitionAsync();
                picture_btn.Image = Properties.Resources.green;
                startToolStripMenuItem.Text = "Start";
                pictureBox1.Enabled = true;                
                isRecognizing = false;
                timer1.Stop();
                timer1.Enabled = false;
            }
            else
            {
                picture_btn.Image = Properties.Resources.red;
                startToolStripMenuItem.Text = "Stop";
                pictureBox1.Enabled = false;
                recognizer.StartContinuousRecognitionAsync();
                isRecognizing = true;
                timer1.Interval = 600;
                timer1.Start();
                timer1.Enabled = true;

            }
        }

        private void pictureBox1_Click(object sender, EventArgs e)
        {
            Startstop();
        }

        private void Form1_Move(object sender, EventArgs e)
        {
            if (this.WindowState == FormWindowState.Normal)
            {
                ShowInTaskbar = true;
                notifyIcon1.Visible = true;
                this.Hide();
                notifyIcon1.ShowBalloonTip(1000);
            }
        }

        private void Form1_MouseDoubleClick(object sender, MouseEventArgs e)
        {
            ShowInTaskbar = true;
            notifyIcon1.Visible = false;
            WindowState = FormWindowState.Normal;
            this.WindowState = FormWindowState.Normal;
            notifyIcon1.Visible = false;

        }

        private void exitToolStripMenuItem_Click(object sender, EventArgs e)
        {
            Application.Exit();
        }


        void SettingFormClosed(object sender, FormClosedEventArgs e)
        {
            initRecognizer();
        }

        private void startToolStripMenuItem_Click(object sender, EventArgs e)
        {
            Startstop();
        }



        private void timer1_Tick(object sender, EventArgs e)
        {
            if (picture_btn.Tag.Equals("red"))
            {
                picture_btn.Image = Properties.Resources.grey;
                picture_btn.Tag = "grey";
            }
            else
            {
                picture_btn.Image = Properties.Resources.red;
                picture_btn.Tag = "red";
            }
        }

        private void pictureBox1_Click_1(object sender, EventArgs e)
        {
            var myForm = new Form2();
            myForm.FormClosed += SettingFormClosed;
            myForm.Show();
        }

        private void notifyIcon1_MouseDoubleClick(object sender, MouseEventArgs e)
        {
            this.Show();
        }

        private void Form1_FormClosing(object sender, FormClosingEventArgs e)
        {
            if (e.CloseReason == CloseReason.UserClosing)
            {
                notifyIcon1.Visible = true;
                this.Hide();
                e.Cancel = true;
            }
        }
    }
}

Answer 1

在深入研究服務之前，您應該始終通讀服務文檔。 該文檔將涵蓋重要的配置方面，並應詳細說明限制。

Azure - 認知服務 - 語音服務 - 語音轉文本文檔
這是語音到文本服務的所有標准資源的登錄頁面，通讀所有這些以了解該服務的工作原理並獲得常見場景的代碼示例。

• 獲取語音識別結果
  這是特定於SpeechRecognizer事件和Results的指南
• 確認的偏移量和持續時間
  本節說明如何獲得詳細結果。

第 1 步 - 配置

您需要配置SpeechRecognizer實例以返回詳細的 output：

SpeechConfig config = SpeechConfig.FromSubscription("key", "region");
// Detailed output will include confidence factor
config.OutputFormat = OutputFormat.Detailed;

if (Properties.Settings.Default.Punctuation)
{
    config.SetServiceProperty("punctuation", "explicit", ServicePropertyChannel.UriQueryParameter);
}

第 2 步 - 從Best()擴展方法訪問詳細信息

這在本節中記錄： 識別偏移量和持續時間但是一般的想法是我們調用e.Result.Best()擴展方法來檢索有關結果的詳細信息。

確保您具有以下 using 語句以使Best()方法可用：

 using Microsoft.CognitiveServices.Speech;

private void SpeechRecognizer_Recognized(object sender, SpeechRecognitionEventArgs e)
{
    if (e.Result.Reason == ResultReason.RecognizedSpeech)
    {
        if (e.Result.Text.ToLower().Equals("new line") || e.Result.Text.ToLower().Equals("newline"))
            SendKeys.SendWait(Environment.NewLine);
        else
            SendKeys.SendWait(e.Result.Text);

        // Get the detailed results
        var detailedResults = e.Result.Best();
        if (detailedResults != null && detailedResults.Any())
        {
            // The first item in detailedResults corresponds to the recognized text.
            // This is not necessarily the item with the highest confidence number.
            var bestResults = detailedResults?.ToList()[0];
            Console.WriteLine(String.Format("\tConfidence: {0}\n\tText: {1}\n\tLexicalForm: {2}\n\tNormalizedForm: {3}\n\tMaskedNormalizedForm: {4}",
                bestResults.Confidence, 
                bestResults.Text, 
                bestResults.LexicalForm, 
                bestResults.NormalizedForm, 
                bestResults.MaskedNormalizedForm));  
    }
}

例如，您可以從服務中獲取此級別的信息，我還啟用了speechConfig.RequestWordLevelTimestamps()以獲取單詞級別的時間戳：