File size: 5,692 Bytes
339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b c9c2be1 339957b a0d4c39 c9c2be1 339957b 45dec12 339957b a0d4c39 339957b a0d4c39 f045571 552dbf1 339957b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 |
from __future__ import absolute_import, division, print_function, unicode_literals
from flask import Flask, make_response, render_template, request, jsonify, redirect, url_for, send_from_directory
import os
import sys
import pytz
import librosa
import shutil
import random
import string
import warnings
import datetime
import librosa.display
import numpy as np
import tensorflow as tf
import gradio as gr
# import pyaudio
# import wave
from tqdm import tqdm
from keras.models import Sequential
from keras.layers import Dense
from keras.utils import to_categorical
from keras.layers import Flatten, Dropout, Activation
from keras.layers import Conv2D, MaxPooling2D
from keras.layers import BatchNormalization
from sklearn.model_selection import train_test_split
from save_data import flag
warnings.filterwarnings("ignore")
timestamp = datetime.datetime.now()
current_date = timestamp.strftime('%d-%m-%Y')
current_time = timestamp.strftime('%I:%M:%S')
IP = ''
cwd = os.getcwd()
classLabels = ('Angry', 'Fear', 'Disgust', 'Happy', 'Sad', 'Surprised', 'Neutral')
numLabels = len(classLabels)
in_shape = (39,216)
model = Sequential()
model.add(Conv2D(8, (13, 13), input_shape=(in_shape[0], in_shape[1], 1)))
model.add(BatchNormalization(axis=-1))
model.add(Activation('relu'))
model.add(Conv2D(8, (13, 13)))
model.add(BatchNormalization(axis=-1))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 1)))
model.add(Conv2D(8, (3, 3)))
model.add(BatchNormalization(axis=-1))
model.add(Activation('relu'))
model.add(Conv2D(8, (1, 1)))
model.add(BatchNormalization(axis=-1))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 1)))
model.add(Flatten())
model.add(Dense(64))
model.add(BatchNormalization())
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(Dense(numLabels, activation='softmax'))
model.compile(loss='binary_crossentropy', optimizer='adam',
metrics=['accuracy'])
# print(model.summary(), file=sys.stderr)
model.load_weights('speech_emotion_detection_ravdess_savee.h5')
# app = Flask(__name__)
# app._static_folder = os.path.join( "/home/ubuntu/Desktop/nlpdemos/server_demos/speech_emotion/static" )
def selected_audio(audio):
try:
if audio and audio != 'Please select any of the following options':
post_file_name = audio.lower() + '.wav'
filepath = os.path.join("pre_recoreded",post_file_name)
if os.path.exists(filepath):
print("SELECT file name => ",filepath)
result = predict_speech_emotion(filepath)
print("result = ",result)
return result
except Exception as e:
print(e)
return "ERROR"
def recorded_audio(audio):
get_audio_name = ''
try:
if audio:
get_audio_name = ''.join([random.choice(string.ascii_letters + string.digits) for n in range(5)])
audio_file_path = audio.name
final_output = predict_speech_emotion(audio_file_path)
flag(audio_file_path,get_audio_name,final_output)
return final_output
except Exception as e:
print(e)
return "ERROR"
def predict_speech_emotion(filepath):
if os.path.exists(filepath):
print("last file name => ",filepath)
X, sample_rate = librosa.load(filepath, res_type='kaiser_best',duration=2.5,sr=22050*2,offset=0.5)
sample_rate = np.array(sample_rate)
mfccs = librosa.feature.mfcc(y=X, sr=sample_rate, n_mfcc=39)
feature = mfccs
feature = feature.reshape(39, 216, 1)
# np_array = np.array([feature])
np_array = np.array([feature])
prediction = model.predict(np_array)
np_argmax = np.argmax(prediction)
result = classLabels[np_argmax]
return result
def return_audio_clip(audio_text):
post_file_name = audio_text.lower() + '.wav'
filepath = os.path.join("pre_recoreded",post_file_name)
return filepath
with gr.Blocks(css=".gradio-container {background-color: lightgray;}") as demo:
gr.Markdown("""<h1 style='text-align: center;>Audio Emotion Detection</h1>""")
with gr.Row():
with gr.Column():
input_audio_text = gr.Dropdown(lable="Input Audio",choices=["Please select any of the following options","Angry", "Happy", "Sad", "Disgust","Fear", "Surprise", "Neutral"],interactive=True)
audio_ui=gr.Audio()
input_audio_text.change(return_audio_clip,input_audio_text,audio_ui)
output_text = gr.Textbox(lable="Prdicted emotion")
sub_btn = gr.Button("Submit")
with gr.Column():
audio=gr.Audio(labele="Recored audio",source="microphone", type="file")
recorded_text = gr.Textbox(lable="Prdicted emotion")
with gr.Column():
sub_btn2 = gr.Button("Submit")
gr.Markdown("""Feel free to give us your [feedback](https://www.pragnakalp.com/contact/) and contact us at [[email protected]]("mailto:[email protected]")
if you want to have your own Speech emotion detection system. We are just one click away. And don't forget to check out more interesting
[NLP services](https://www.pragnakalp.com/services/natural-language-processing-services/) we are offering.
<p style='text-align: center;'>Developed by :<a href="https://www.pragnakalp.com" target="_blank"> Pragnakalp Techlabs</a></p>""")
sub_btn.click(selected_audio, inputs=input_audio_text, outputs=output_text)
sub_btn2.click(recorded_audio, inputs=audio, outputs=recorded_text)
demo.launch() |