def find_music(audio_file):
modelName = "pyAA/data/svmSM"
[Fs, x] = aIO.readAudioFile(audio_file)
duration = x.shape[0] / float(Fs)
t1 = time.clock()
flagsInd, classNames, acc, CMt = aS.mtFileClassification(
audio_file, modelName, "svm", False, '')
[
Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT
] = aT.loadSVModel(modelName)
t2 = time.clock()
perTime1 = duration / (t2 - t1)
flags = [classNames[int(f)] for f in flagsInd]
(segs, classes) = aS.flags2segs(flags, mtStep)
i = 0 #len(classes)-1
file_parts = []
cbn = sox.Combiner()
if len(classes) > 1:
for c in classes:
if c == 'music':
start = segs[i][0]
if i != 0:
start -= 0.5
end = segs[i][1]
if i != len(classes) - 1:
end += 2.5
file_parts.append((int(start * 1000), int(end * 1000)))
i += 1
return file_parts
def initSubscriber():
global features_subscriber
global modelName
global classifierInfo
global classification_publisher
rospy.init_node("audio_features_classifier_node")
modelName = rospy.get_param('~classifier_name', 'modelSVM')
sub_topic = rospy.get_param('~features_topic','/audio_features_extraction/features')
pub_topic = rospy.get_param('~classification_topic','~audio_classification')
[Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep, computeBEAT] = audioTrainTest.loadSVModel(os.path.dirname(os.path.realpath(sys.argv[0]))+'/classifier_data/'+modelName)
classifierInfo["Classifier"] = Classifier
classifierInfo["MEAN"] = MEAN
classifierInfo["STD"] = STD
classifierInfo["classNames"] = classNames
classifierInfo["mtWin"] = mtWin
classifierInfo["mtStep"] = mtStep
classifierInfo["stWin"] = stWin
classifierInfo["stStep"] = stStep
classifierInfo["computeBEAT"] = computeBEAT
print MEAN
classification_publisher = rospy.Publisher("~"+pub_topic, classificationResult, queue_size=10)
features_subscriber = rospy.Subscriber(sub_topic, featMsg, featuresCallback)
print "Waiting for features_topic to be published..."
rospy.spin()
def __init__(self, model_path=None):
assert (self.model_type == 'svm')
if not model_path:
ppath = os.path.join(os.environ['HOME'], '.speechworker')
model_path = os.path.join(ppath, 'models/svmSM')
[
self.classifier, self.model_mean, self.model_sd, self.class_names,
self.mt_win, self.mt_step, self.st_win, self.st_step, _
] = aT.loadSVModel(model_path)
def getMusicSegmentsFromFile(inputFile):
modelType = "svm"
modelName = "data/svmMovies8classes"
dirOutput = inputFile[0:-4] + "_musicSegments"
if os.path.exists(dirOutput) and dirOutput != ".":
shutil.rmtree(dirOutput)
os.makedirs(dirOutput)
[Fs, x] = audioBasicIO.readAudioFile(inputFile)
if modelType == 'svm':
[
Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT
] = aT.loadSVModel(modelName)
elif modelType == 'knn':
[
Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT
] = aT.loadKNNModel(modelName)
flagsInd, classNames, acc, CM = aS.mtFileClassification(inputFile,
modelName,
modelType,
plotResults=False,
gtFile="")
segs, classes = aS.flags2segs(flagsInd, mtStep)
for i, s in enumerate(segs):
if (classNames[int(classes[i])]
== "Music") and (s[1] - s[0] >= minDuration):
strOut = "{0:s}{1:.3f}-{2:.3f}.wav".format(dirOutput + os.sep,
s[0], s[1])
wavfile.write(strOut, Fs, x[int(Fs * s[0]):int(Fs * s[1])])
def mtFileClassification(inputFile, modelName, modelType, plotResults=False, gtFile=""):
'''
This function performs mid-term classification of an audio stream.
Towards this end, supervised knowledge is used, i.e. a pre-trained classifier.
ARGUMENTS:
- inputFile: path of the input WAV file
- modelName: name of the classification model
- modelType: svm or knn depending on the classifier type
- plotResults: True if results are to be plotted using matplotlib along with a set of statistics
RETURNS:
- segs: a sequence of segment's endpoints: segs[i] is the endpoint of the i-th segment (in seconds)
- classes: a sequence of class flags: class[i] is the class ID of the i-th segment
'''
if not os.path.isfile(modelName):
print("mtFileClassificationError: input modelType not found!")
return (-1, -1, -1, -1)
# Load classifier:
if (modelType == 'svm') or (modelType == 'svm_rbf'):
[Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin,
stStep, computeBEAT] = aT.loadSVModel(modelName)
elif modelType == 'knn':
[Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin,
stStep, computeBEAT] = aT.loadKNNModel(modelName)
elif modelType == 'randomforest':
[Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin,
stStep, computeBEAT] = aT.loadRandomForestModel(modelName)
elif modelType == 'gradientboosting':
[Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT] = aT.loadGradientBoostingModel(modelName)
elif modelType == 'extratrees':
[Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin,
stStep, computeBEAT] = aT.loadExtraTreesModel(modelName)
if computeBEAT:
print("Model " + modelName +
" contains long-term music features (beat etc) and cannot be used in segmentation")
return (-1, -1, -1, -1)
[Fs, x] = audioBasicIO.readAudioFile(inputFile) # load input file
if Fs == -1: # could not read file
return (-1, -1, -1, -1)
# convert stereo (if) to mono
x = audioBasicIO.stereo2mono(x)
Duration = len(x) / Fs
# mid-term feature extraction:
[MidTermFeatures, _] = aF.mtFeatureExtraction(
x, Fs, mtWin * Fs, mtStep * Fs, round(Fs * stWin), round(Fs * stStep))
flags = []
Ps = []
flagsInd = []
# for each feature vector (i.e. for each fix-sized segment):
for i in range(MidTermFeatures.shape[1]):
# normalize current feature vector
curFV = (MidTermFeatures[:, i] - MEAN) / STD
[Result, P] = aT.classifierWrapper(
Classifier, modelType, curFV) # classify vector
flagsInd.append(Result)
# update class label matrix
flags.append(classNames[int(Result)])
# update probability matrix
Ps.append(numpy.max(P))
flagsInd = numpy.array(flagsInd)
# 1-window smoothing
for i in range(1, len(flagsInd) - 1):
if flagsInd[i - 1] == flagsInd[i + 1]:
flagsInd[i] = flagsInd[i + 1]
# convert fix-sized flags to segments and classes
(segs, classes) = flags2segs(flags, mtStep)
segs[-1] = len(x) / float(Fs)
# Load grount-truth:
if os.path.isfile(gtFile):
[segStartGT, segEndGT, segLabelsGT] = readSegmentGT(gtFile)
flagsGT, classNamesGT = segs2flags(
segStartGT, segEndGT, segLabelsGT, mtStep)
flagsIndGT = []
for j, fl in enumerate(flagsGT): # "align" labels with GT
if classNamesGT[flagsGT[j]] in classNames:
flagsIndGT.append(classNames.index(classNamesGT[flagsGT[j]]))
else:
flagsIndGT.append(-1)
flagsIndGT = numpy.array(flagsIndGT)
CM = numpy.zeros((len(classNamesGT), len(classNamesGT)))
for i in range(min(flagsInd.shape[0], flagsIndGT.shape[0])):
CM[int(flagsIndGT[i]), int(flagsInd[i])] += 1
else:
CM = []
flagsIndGT = numpy.array([])
acc = plotSegmentationResults(
flagsInd, flagsIndGT, classNames, mtStep, not plotResults)
if acc >= 0:
print("Overall Accuracy: {0:.3f}".format(acc))
return (flagsInd, classNamesGT, acc, CM)
else:
return (flagsInd, classNames, acc, CM)
def recordAnalyzeAudio(duration, outputWavFile, midTermBufferSizeSec,
modelName, modelType):
'''
recordAnalyzeAudio(duration, outputWavFile, midTermBufferSizeSec, modelName, modelType)
This function is used to record and analyze audio segments, in a fix window basis.
ARGUMENTS:
- duration total recording duration
- outputWavFile path of the output WAV file
- midTermBufferSizeSec (fix)segment length in seconds
- modelName classification model name
- modelType classification model type
'''
if modelType == 'svm':
[
Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT
] = aT.loadSVModel(modelName)
elif modelType == 'knn':
[
Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT
] = aT.loadKNNModel(modelName)
else:
Classifier = None
inp = alsaaudio.PCM(alsaaudio.PCM_CAPTURE, alsaaudio.PCM_NONBLOCK)
inp.setchannels(1)
inp.setrate(Fs)
inp.setformat(alsaaudio.PCM_FORMAT_S16_LE)
inp.setperiodsize(512)
midTermBufferSize = int(midTermBufferSizeSec * Fs)
allData = []
midTermBuffer = []
curWindow = []
count = 0
while len(allData) < duration * Fs:
# Read data from device
l, data = inp.read()
if l:
for i in range(l):
curWindow.append(audioop.getsample(data, 2, i))
if (len(curWindow) + len(midTermBuffer) > midTermBufferSize):
samplesToCopyToMidBuffer = midTermBufferSize - \
len(midTermBuffer)
else:
samplesToCopyToMidBuffer = len(curWindow)
midTermBuffer = midTermBuffer + \
curWindow[0:samplesToCopyToMidBuffer]
del (curWindow[0:samplesToCopyToMidBuffer])
if len(midTermBuffer) == midTermBufferSize:
count += 1
if Classifier != None:
[mtFeatures,
stFeatures] = aF.mtFeatureExtraction(midTermBuffer, Fs,
2.0 * Fs, 2.0 * Fs,
0.020 * Fs, 0.020 * Fs)
curFV = (mtFeatures[:, 0] - MEAN) / STD
[result, P] = aT.classifierWrapper(Classifier, modelType,
curFV)
print(classNames[int(result)])
allData = allData + midTermBuffer
plt.clf()
plt.plot(midTermBuffer)
plt.show(block=False)
plt.draw()
midTermBuffer = []
allDataArray = numpy.int16(allData)
wavfile.write(outputWavFile, Fs, allDataArray)
def classifyFolderWrapper(inputFolder, modelType, modelName, outputMode=False):
if not os.path.isfile(modelName):
raise Exception("Input modelName not found!")
if modelType == 'svm':
[
Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT
] = aT.loadSVModel(modelName)
elif modelType == 'knn':
[
Classifier, MEAN, STD, classNames, mtWin, mtStep, stWin, stStep,
computeBEAT
] = aT.loadKNNModel(modelName)
PsAll = numpy.zeros((len(classNames), ))
files = "*.wav"
if os.path.isdir(inputFolder):
strFilePattern = os.path.join(inputFolder, files)
else:
strFilePattern = inputFolder + files
wavFilesList = []
wavFilesList.extend(glob.glob(strFilePattern))
wavFilesList = sorted(wavFilesList)
if len(wavFilesList) == 0:
print("No WAV files found!")
return
Results = []
for wavFile in wavFilesList:
[Fs, x] = audioBasicIO.readAudioFile(wavFile)
signalLength = x.shape[0] / float(Fs)
[Result, P,
classNames] = aT.fileClassification(wavFile, modelName, modelType)
PsAll += (numpy.array(P) * signalLength)
Result = int(Result)
Results.append(Result)
if outputMode:
print("{0:s}\t{1:s}".format(wavFile, classNames[Result]))
Results = numpy.array(Results)
# print distribution of classes:
[Histogram, _] = numpy.histogram(Results,
bins=numpy.arange(len(classNames) + 1))
if outputMode:
for i, h in enumerate(Histogram):
print("{0:20s}\t\t{1:d}".format(classNames[i], h))
PsAll = PsAll / numpy.sum(PsAll)
if outputMode:
fig = plt.figure()
ax = fig.add_subplot(111)
plt.title("Classes percentage " + inputFolder.replace('Segments', ''))
ax.axis((0, len(classNames) + 1, 0, 1))
ax.set_xticks(numpy.array(range(len(classNames) + 1)))
ax.set_xticklabels([" "] + classNames)
ax.bar(numpy.array(range(len(classNames))) + 0.5, PsAll)
plt.show()
return classNames, PsAll