PocketSphinx安卓显示结果，即使没有我说话

Question

我在android上使用pocketsphinx来查找关键字，但它不识别除所需关键字之外的任何其他关键字。此外，它甚至不等待我发言，并在logcat中显示关键字。

这是我的代码：

public class SpeechService extends Service implements RecognitionListener {
private static final String TAG = "myTag";
private static final String KWS_SEARCH = "wakeup";
private static final String KEY_PHRASE = "hello";
private SpeechRecognizer recognizer;
private PowerManager.WakeLock lock;
@Override
public IBinder onBind(Intent intent) {
    return null;
}
@Override
public void onDestroy() {
    super.onDestroy();
    recognizer.cancel();
    recognizer.shutdown();
    lock.release();
}
@Override
public int onStartCommand(Intent data, int flags, int startId) {
    new AsyncTask<Void, Void, Exception>() {
        @Override
        protected Exception doInBackground(Void... params) {
            try {
                Assets assets = new Assets(SpeechService.this);
                File assetDir = assets.syncAssets();
                setupRecognizer(assetDir);
            } catch (IOException e) {
                return e;
            }
            return null;
        }
        @Override
        protected void onPostExecute(Exception result) {
            if(result == null)
                recognizer.startListening(KWS_SEARCH);
        }
    }.execute();
    return START_STICKY;
}
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-40f)

            .setBoolean("-allphone_ci", true)
            .getRecognizer();
    recognizer.addListener(this);
    // Create keyword-activation search.
    recognizer.addKeyphraseSearch(KWS_SEARCH, KEY_PHRASE);
}
@Override
public void onBeginningOfSpeech() {
    //Log.i(TAG, "onBeginningOfSpeech");
}
@Override
public void onEndOfSpeech() {
    //Log.i(TAG, "onEndOfSpeech");
    if(!recognizer.getSearchName().equals(KWS_SEARCH)) {
        recognizer.cancel();
        recognizer.startListening(KWS_SEARCH);
    }
}
@Override
public void onPartialResult(Hypothesis hypothesis) {
    //Log.i(TAG, "onPartialResult");
    if(hypothesis == null)
        return;
    String result = hypothesis.getHypstr();
    Log.i(TAG, "result = " + result);
    recognizer.cancel();
    recognizer.startListening(KWS_SEARCH);
}
@Override
public void onResult(Hypothesis hypothesis) {
    //Log.i(TAG, "onResult");
}
@Override
public void onError(Exception e) {
    Log.i(TAG, "onError");
    recognizer.startListening(KWS_SEARCH);
}
@Override
public void onTimeout() {
    Log.i(TAG, "onTimeout");
    recognizer.startListening(KWS_SEARCH);
}
}

这是我还没说话的时候的Logcat。

result = hello
result = hello
result = hello hello
result = hello
result = hello hello
result = hello

等等…

我做错了什么？？？

Answer 1

如果您有太多的假警报，您可以将关键字阈值从1e-40降低到1e-20，甚至1e-10：

        .setKeywordThreshold(1e-40f)

Answer 2

我也遇到了同样的问题。

我觉得识别"你好"太容易了，所以我第一时间把关键词改成了两个词。

但它使用起来并不好，因为有时它无法识别。

所以我自己做了语法，并使用了addGrammarSearch方法。顺便说一句，你的语法不能只有一个单词，你必须添加一些单词来欺骗噪音。

最后我写了判断法来识别onParticalResult或onResult方法中的关键字。

如果你有更好的答案，我想知道。