使用PocketSphinx识别多个关键字

Question

我已经安装了PocketSphinx演示程序，并且在Ubuntu和Eclipse下运行良好，但是尽管尝试了一下，但仍无法弄清楚如何添加多个单词的识别。我只想让代码识别单个单词，然后我就可以在代码中识别它们switch()，例如“上”，“下”，“左”，“右”。我不想识别句子，只能识别单个单词。任何帮助，将不胜感激。我发现其他用户也有类似的问题，但到目前为止，没人知道答案。让我感到困惑的一件事是，为什么我们根本需要使用“唤醒”常量？private static final String KWS_SEARCH = "wakeup";private static final String KEYPHRASE = "oh mighty computer";...recognizer.addKeyphraseSearch(KWS_SEARCH, KEYPHRASE);有wakeup什么关系吗？我已经取得了一些进步（？）：使用，addGrammarSearch我可以使用.gram文件列出我的单词，例如up,down,left,right,forwards,backwards，如果我只说那些特定的单词，这似乎很好用。但是，任何其他字词都将导致系统将所陈述的内容与“最近”字词进行匹配。理想情况下，如果.gram文件中没有说出的单词，我不希望被识别...

Answer 1

您可以使用addKeywordSearch哪个用于归档密钥短语。每行一个短语，例如//中的每个短语都有阈值

up /1.0/

down /1.0/

left /1.0/

right /1.0/

forwards /1e-1/

必须选择阈值以避免错误警报。

Answer 2

正在为PocketSphinx演示更新Antinous修订，以使其能够在Android Studio上运行。这就是我到目前为止

//Note: change MainActivity to PocketSphinxActivity for demo use...

public class MainActivity extends Activity implements RecognitionListener {

private static final String DIGITS_SEARCH = "digits";

private SpeechRecognizer recognizer;

/* Used to handle permission request */

private static final int PERMISSIONS_REQUEST_RECORD_AUDIO = 1;

@Override

public void onCreate(Bundle state) {

    super.onCreate(state);

    setContentView(R.layout.main);

    ((TextView) findViewById(R.id.caption_text))

            .setText("Preparing the recognizer");

    // Check if user has given permission to record audio

    int permissionCheck = ContextCompat.checkSelfPermission(getApplicationContext(), Manifest.permission.RECORD_AUDIO);

    if (permissionCheck != PackageManager.PERMISSION_GRANTED) {

        ActivityCompat.requestPermissions(this, new String[]{Manifest.permission.RECORD_AUDIO}, PERMISSIONS_REQUEST_RECORD_AUDIO);

        return;

    }

    new AsyncTask<Void, Void, Exception>() {

        @Override

        protected Exception doInBackground(Void... params) {

            try {

                Assets assets = new Assets(MainActivity.this);

                File assetDir = assets.syncAssets();

                setupRecognizer(assetDir);

            } catch (IOException e) {

                return e;

            }

            return null;

        }

        @Override

        protected void onPostExecute(Exception result) {

            if (result != null) {

                ((TextView) findViewById(R.id.caption_text))

                        .setText("Failed to init recognizer " + result);

            } else {

                reset();

            }

        }

    }.execute();

    ((TextView) findViewById(R.id.caption_text)).setText("Say one, two, three, four, five, six...");

}

/**

 * In partial result we get quick updates about current hypothesis. In

 * keyword spotting mode we can react here, in other modes we need to wait

 * for final result in onResult.

 */

@Override

public void onPartialResult(Hypothesis hypothesis) {

    if (hypothesis == null) {

        return;

    } else if (hypothesis != null) {

        if (recognizer != null) {

            //recognizer.rapidSphinxPartialResult(hypothesis.getHypstr());

            String text = hypothesis.getHypstr();

            if (text.equals(DIGITS_SEARCH)) {

                recognizer.cancel();

                performAction();

                recognizer.startListening(DIGITS_SEARCH);

            }else{

                //Toast.makeText(getApplicationContext(),"Partial result = " +text,Toast.LENGTH_SHORT).show();

            }

        }

    }

}

@Override

public void onResult(Hypothesis hypothesis) {

    ((TextView) findViewById(R.id.result_text)).setText("");

    if (hypothesis != null) {

        String text = hypothesis.getHypstr();

        makeText(getApplicationContext(), "Hypothesis" +text, Toast.LENGTH_SHORT).show();

    }else if(hypothesis == null){

        makeText(getApplicationContext(), "hypothesis = null", Toast.LENGTH_SHORT).show();

    }

}

@Override

public void onDestroy() {

    super.onDestroy();

    recognizer.cancel();

    recognizer.shutdown();

}

@Override

public void onBeginningOfSpeech() {

}

@Override

public void onEndOfSpeech() {

   reset();

}

@Override

public void onTimeout() {

}

private void setupRecognizer(File assetsDir) throws IOException {

    // The recognizer can be configured to perform multiple searches

    // of different kind and switch between them

    recognizer = defaultSetup()

            .setAcousticModel(new File(assetsDir, "en-us-ptm"))

            .setDictionary(new File(assetsDir, "cmudict-en-us.dict"))

            // .setRawLogDir(assetsDir).setKeywordThreshold(1e-20f)

            .getRecognizer();

    recognizer.addListener(this);

    File digitsGrammar = new File(assetsDir, "digits.gram");

    recognizer.addKeywordSearch(DIGITS_SEARCH, digitsGrammar);

}

private void reset(){

    recognizer.stop();

    recognizer.startListening(DIGITS_SEARCH);

}

@Override

public void onError(Exception error) {

    ((TextView) findViewById(R.id.caption_text)).setText(error.getMessage());

}

public void performAction() {

    // do here whatever you want

    makeText(getApplicationContext(), "performAction done... ", Toast.LENGTH_SHORT).show();

}

}

请注意：此工作正在进行中。过一会再来检查。建议将不胜感激。