Easy Audio Classification with Keras

0/5 No votes

Report this app

Description

[ad_1]

Introduction

On this tutorial we are going to construct a deep studying mannequin to categorise phrases. We are going to use tfdatasets to deal with information IO and pre-processing, and Keras to construct and prepare the mannequin.

We are going to use the Speech Instructions dataset which consists of 65,000 one-second audio recordsdata of individuals saying 30 completely different phrases. Every file comprises a single spoken English phrase. The dataset was launched by Google underneath CC License.

Our mannequin is a Keras port of the TensorFlow tutorial on Easy Audio Recognition which in flip was impressed by Convolutional Neural Networks for Small-footprint Key phrase Recognizing. There are different approaches to the speech recognition process, like recurrent neural networks, dilated (atrous) convolutions or Studying from Between-class Examples for Deep Sound Recognition.

The mannequin we are going to implement right here isn’t the state-of-the-art for audio recognition techniques, that are far more complicated, however is comparatively easy and quick to coach. Plus, we present learn how to effectively use tfdatasets to preprocess and serve information.

Audio illustration

Many deep studying fashions are end-to-end, i.e. we let the mannequin be taught helpful representations instantly from the uncooked information. Nevertheless, audio information grows very quick – 16,000 samples per second with a really wealthy construction at many time-scales. With a purpose to keep away from having to cope with uncooked wave sound information, researchers normally use some form of characteristic engineering.

Each sound wave might be represented by its spectrum, and digitally it may be computed utilizing the Quick Fourier Remodel (FFT).

A standard solution to symbolize audio information is to interrupt it into small chunks, which normally overlap. For every chunk we use the FFT to calculate the magnitude of the frequency spectrum. The spectra are then mixed, facet by facet, to type what we name a spectrogram.

It’s additionally frequent for speech recognition techniques to additional rework the spectrum and compute the Mel-Frequency Cepstral Coefficients. This transformation takes under consideration that the human ear can’t discern the distinction between two carefully spaced frequencies and neatly creates bins on the frequency axis. A fantastic tutorial on MFCCs might be discovered right here.

By Aquegg - Own work, Public Domain, https://commons.wikimedia.org/w/index.php?curid=5544473

After this process, now we have a picture for every audio pattern and we are able to use convolutional neural networks, the usual structure kind in picture recognition fashions.

Downloading

First, let’s obtain information to a listing in our undertaking. You may both obtain from this hyperlink (~1GB) or from R with:

dir.create("information")

obtain.file(
  url = "http://obtain.tensorflow.org/information/speech_commands_v0.01.tar.gz", 
  destfile = "information/speech_commands_v0.01.tar.gz"
)

untar("information/speech_commands_v0.01.tar.gz", exdir = "information/speech_commands_v0.01")

Contained in the information listing we could have a folder referred to as speech_commands_v0.01. The WAV audio recordsdata inside this listing are organised in sub-folders with the label names. For instance, all one-second audio recordsdata of individuals talking the phrase “mattress” are contained in the mattress listing. There are 30 of them and a particular one referred to as _background_noise_ which comprises varied patterns that could possibly be blended in to simulate background noise.

Importing

On this step we are going to listing all audio .wav recordsdata right into a tibble with 3 columns:

  • fname: the file identify;
  • class: the label for every audio file;
  • class_id: a singular integer quantity ranging from zero for every class – used to one-hot encode the courses.

This shall be helpful to the subsequent step after we will create a generator utilizing the tfdatasets bundle.

Generator

We are going to now create our Dataset, which within the context of tfdatasets, provides operations to the TensorFlow graph with a view to learn and pre-process information. Since they’re TensorFlow ops, they’re executed in C++ and in parallel with mannequin coaching.

The generator we are going to create shall be chargeable for studying the audio recordsdata from disk, creating the spectrogram for each and batching the outputs.

Let’s begin by creating the dataset from slices of the information.body with audio file names and courses we simply created.

Now, let’s outline the parameters for spectrogram creation. We have to outline window_size_ms which is the dimensions in milliseconds of every chunk we are going to break the audio wave into, and window_stride_ms, the space between the facilities of adjoining chunks:

window_size_ms <- 30
window_stride_ms <- 10

Now we are going to convert the window measurement and stride from milliseconds to samples. We’re contemplating that our audio recordsdata have 16,000 samples per second (1000 ms).

window_size <- as.integer(16000*window_size_ms/1000)
stride <- as.integer(16000*window_stride_ms/1000)

We are going to get hold of different portions that shall be helpful for spectrogram creation, just like the variety of chunks and the FFT measurement, i.e., the variety of bins on the frequency axis. The perform we’re going to use to compute the spectrogram doesn’t permit us to vary the FFT measurement and as a substitute by default makes use of the primary energy of two larger than the window measurement.

We are going to now use dataset_map which permits us to specify a pre-processing perform for every statement (line) of our dataset. It’s on this step that we learn the uncooked audio file from disk and create its spectrogram and the one-hot encoded response vector.

# shortcuts to used TensorFlow modules.
audio_ops <- tf$contrib$framework$python$ops$audio_ops

ds <- ds %>%
  dataset_map(perform(obs) {
    
    # a great way to debug when constructing tfdatsets pipelines is to make use of a print
    # assertion like this:
    # print(str(obs))
    
    # decoding wav recordsdata
    audio_binary <- tf$read_file(tf$reshape(obs$fname, form = listing()))
    wav <- audio_ops$decode_wav(audio_binary, desired_channels = 1)
    
    # create the spectrogram
    spectrogram <- audio_ops$audio_spectrogram(
      wav$audio, 
      window_size = window_size, 
      stride = stride,
      magnitude_squared = TRUE
    )
    
    # normalization
    spectrogram <- tf$log(tf$abs(spectrogram) + 0.01)
    
    # transferring channels to final dim
    spectrogram <- tf$transpose(spectrogram, perm = c(1L, 2L, 0L))
    
    # rework the class_id right into a one-hot encoded vector
    response <- tf$one_hot(obs$class_id, 30L)
    
    listing(spectrogram, response)
  }) 

Now, we are going to specify how we would like batch observations from the dataset. We’re utilizing dataset_shuffle since we need to shuffle observations from the dataset, in any other case it might observe the order of the df object. Then we use dataset_repeat with a view to inform TensorFlow that we need to maintain taking observations from the dataset even when all observations have already been used. And most significantly right here, we use dataset_padded_batch to specify that we would like batches of measurement 32, however they need to be padded, ie. if some statement has a special measurement we pad it with zeroes. The padded form is handed to dataset_padded_batch by way of the padded_shapes argument and we use NULL to state that this dimension doesn’t should be padded.

ds <- ds %>% 
  dataset_shuffle(buffer_size = 100) %>%
  dataset_repeat() %>%
  dataset_padded_batch(
    batch_size = 32, 
    padded_shapes = listing(
      form(n_chunks, fft_size, NULL), 
      form(NULL)
    )
  )

That is our dataset specification, however we would want to rewrite all of the code for the validation information, so it’s good observe to wrap this right into a perform of the information and different vital parameters like window_size_ms and window_stride_ms. Beneath, we are going to outline a perform referred to as data_generator that may create the generator relying on these inputs.

data_generator <- perform(df, batch_size, shuffle = TRUE, 
                           window_size_ms = 30, window_stride_ms = 10) {
  
  window_size <- as.integer(16000*window_size_ms/1000)
  stride <- as.integer(16000*window_stride_ms/1000)
  fft_size <- as.integer(2^trunc(log(window_size, 2)) + 1)
  n_chunks <- size(seq(window_size/2, 16000 - window_size/2, stride))
  
  ds <- tensor_slices_dataset(df)
  
  if (shuffle) 
    ds <- ds %>% dataset_shuffle(buffer_size = 100)  
  
  ds <- ds %>%
    dataset_map(perform(obs) {
      
      # decoding wav recordsdata
      audio_binary <- tf$read_file(tf$reshape(obs$fname, form = listing()))
      wav <- audio_ops$decode_wav(audio_binary, desired_channels = 1)
      
      # create the spectrogram
      spectrogram <- audio_ops$audio_spectrogram(
        wav$audio, 
        window_size = window_size, 
        stride = stride,
        magnitude_squared = TRUE
      )
      
      spectrogram <- tf$log(tf$abs(spectrogram) + 0.01)
      spectrogram <- tf$transpose(spectrogram, perm = c(1L, 2L, 0L))
      
      # rework the class_id right into a one-hot encoded vector
      response <- tf$one_hot(obs$class_id, 30L)
      
      listing(spectrogram, response)
    }) %>%
    dataset_repeat()
  
  ds <- ds %>% 
    dataset_padded_batch(batch_size, listing(form(n_chunks, fft_size, NULL), form(NULL)))
  
  ds
}

Now, we are able to outline coaching and validation information mills. It’s value noting that executing this received’t truly compute any spectrogram or learn any file. It’s going to solely outline within the TensorFlow graph the way it ought to learn and pre-process information.

set.seed(6)
id_train <- pattern(nrow(df), measurement = 0.7*nrow(df))

ds_train <- data_generator(
  df[id_train,], 
  batch_size = 32, 
  window_size_ms = 30, 
  window_stride_ms = 10
)
ds_validation <- data_generator(
  df[-id_train,], 
  batch_size = 32, 
  shuffle = FALSE, 
  window_size_ms = 30, 
  window_stride_ms = 10
)

To truly get a batch from the generator we might create a TensorFlow session and ask it to run the generator. For instance:

sess <- tf$Session()
batch <- next_batch(ds_train)
str(sess$run(batch))
Record of two
 $ : num [1:32, 1:98, 1:257, 1] -4.6 -4.6 -4.61 -4.6 -4.6 ...
 $ : num [1:32, 1:30] 0 0 0 0 0 0 0 0 0 0 ...

Every time you run sess$run(batch) it’s best to see a special batch of observations.

Mannequin definition

Now that we all know how we are going to feed our information we are able to deal with the mannequin definition. The spectrogram might be handled like a picture, so architectures which might be generally utilized in picture recognition duties ought to work nicely with the spectrograms too.

We are going to construct a convolutional neural community much like what now we have constructed right here for the MNIST dataset.

The enter measurement is outlined by the variety of chunks and the FFT measurement. Like we defined earlier, they are often obtained from the window_size_ms and window_stride_ms used to generate the spectrogram.

We are going to now outline our mannequin utilizing the Keras sequential API:

mannequin <- keras_model_sequential()
mannequin %>%  
  layer_conv_2d(input_shape = c(n_chunks, fft_size, 1), 
                filters = 32, kernel_size = c(3,3), activation = 'relu') %>% 
  layer_max_pooling_2d(pool_size = c(2, 2)) %>% 
  layer_conv_2d(filters = 64, kernel_size = c(3,3), activation = 'relu') %>% 
  layer_max_pooling_2d(pool_size = c(2, 2)) %>% 
  layer_conv_2d(filters = 128, kernel_size = c(3,3), activation = 'relu') %>% 
  layer_max_pooling_2d(pool_size = c(2, 2)) %>% 
  layer_conv_2d(filters = 256, kernel_size = c(3,3), activation = 'relu') %>% 
  layer_max_pooling_2d(pool_size = c(2, 2)) %>% 
  layer_dropout(price = 0.25) %>% 
  layer_flatten() %>% 
  layer_dense(items = 128, activation = 'relu') %>% 
  layer_dropout(price = 0.5) %>% 
  layer_dense(items = 30, activation = 'softmax')

We used 4 layers of convolutions mixed with max pooling layers to extract options from the spectrogram photographs and a couple of dense layers on the high. Our community is relatively easy when in comparison with extra superior architectures like ResNet or DenseNet that carry out very nicely on picture recognition duties.

Now let’s compile our mannequin. We are going to use categorical cross entropy because the loss perform and use the Adadelta optimizer. It’s additionally right here that we outline that we are going to take a look at the accuracy metric throughout coaching.

mannequin %>% compile(
  loss = loss_categorical_crossentropy,
  optimizer = optimizer_adadelta(),
  metrics = c('accuracy')
)

Mannequin becoming

Now, we are going to match our mannequin. In Keras we are able to use TensorFlow Datasets as inputs to the fit_generator perform and we are going to do it right here.

mannequin %>% fit_generator(
  generator = ds_train,
  steps_per_epoch = 0.7*nrow(df)/32,
  epochs = 10, 
  validation_data = ds_validation, 
  validation_steps = 0.3*nrow(df)/32
)
Epoch 1/10
1415/1415 [==============================] - 87s 62ms/step - loss: 2.0225 - acc: 0.4184 - val_loss: 0.7855 - val_acc: 0.7907
Epoch 2/10
1415/1415 [==============================] - 75s 53ms/step - loss: 0.8781 - acc: 0.7432 - val_loss: 0.4522 - val_acc: 0.8704
Epoch 3/10
1415/1415 [==============================] - 75s 53ms/step - loss: 0.6196 - acc: 0.8190 - val_loss: 0.3513 - val_acc: 0.9006
Epoch 4/10
1415/1415 [==============================] - 75s 53ms/step - loss: 0.4958 - acc: 0.8543 - val_loss: 0.3130 - val_acc: 0.9117
Epoch 5/10
1415/1415 [==============================] - 75s 53ms/step - loss: 0.4282 - acc: 0.8754 - val_loss: 0.2866 - val_acc: 0.9213
Epoch 6/10
1415/1415 [==============================] - 76s 53ms/step - loss: 0.3852 - acc: 0.8885 - val_loss: 0.2732 - val_acc: 0.9252
Epoch 7/10
1415/1415 [==============================] - 75s 53ms/step - loss: 0.3566 - acc: 0.8991 - val_loss: 0.2700 - val_acc: 0.9269
Epoch 8/10
1415/1415 [==============================] - 76s 54ms/step - loss: 0.3364 - acc: 0.9045 - val_loss: 0.2573 - val_acc: 0.9284
Epoch 9/10
1415/1415 [==============================] - 76s 53ms/step - loss: 0.3220 - acc: 0.9087 - val_loss: 0.2537 - val_acc: 0.9323
Epoch 10/10
1415/1415 [==============================] - 76s 54ms/step - loss: 0.2997 - acc: 0.9150 - val_loss: 0.2582 - val_acc: 0.9323

The mannequin’s accuracy is 93.23%. Let’s learn to make predictions and try the confusion matrix.

Making predictions

We will use thepredict_generator perform to make predictions on a brand new dataset. Let’s make predictions for our validation dataset. The predict_generator perform wants a step argument which is the variety of occasions the generator shall be referred to as.

We will calculate the variety of steps by realizing the batch measurement, and the dimensions of the validation dataset.

df_validation <- df[-id_train,]
n_steps <- nrow(df_validation)/32 + 1

We will then use the predict_generator perform:

predictions <- predict_generator(
  mannequin, 
  ds_validation, 
  steps = n_steps
  )
str(predictions)
num [1:19424, 1:30] 1.22e-13 7.30e-19 5.29e-10 6.66e-22 1.12e-17 ...

This can output a matrix with 30 columns – one for every phrase and n_steps*batch_size variety of rows. Be aware that it begins repeating the dataset on the finish to create a full batch.

We will compute the expected class by taking the column with the best chance, for instance.

courses <- apply(predictions, 1, which.max) - 1

A pleasant visualization of the confusion matrix is to create an alluvial diagram:

library(dplyr)
library(alluvial)
x <- df_validation %>%
  mutate(pred_class_id = head(courses, nrow(df_validation))) %>%
  left_join(
    df_validation %>% distinct(class_id, class) %>% rename(pred_class = class),
    by = c("pred_class_id" = "class_id")
  ) %>%
  mutate(appropriate = pred_class == class) %>%
  depend(pred_class, class, appropriate)

alluvial(
  x %>% choose(class, pred_class),
  freq = x$n,
  col = ifelse(x$appropriate, "lightblue", "pink"),
  border = ifelse(x$appropriate, "lightblue", "pink"),
  alpha = 0.6,
  cover = x$n < 20
)

We will see from the diagram that essentially the most related mistake our mannequin makes is to categorise “tree” as “three”. There are different frequent errors like classifying “go” as “no”, “up” as “off”. At 93% accuracy for 30 courses, and contemplating the errors we are able to say that this mannequin is fairly cheap.

The saved mannequin occupies 25Mb of disk house, which is affordable for a desktop however might not be on small gadgets. We might prepare a smaller mannequin, with fewer layers, and see how a lot the efficiency decreases.

In speech recognition duties its additionally frequent to do some form of information augmentation by mixing a background noise to the spoken audio, making it extra helpful for actual functions the place it’s frequent to produce other irrelevant sounds taking place within the surroundings.

The total code to breed this tutorial is out there right here.

[ad_2]

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.