Upload 3 files

Voice_Distinction.py

@@ -0,0 +1,100 @@
+# type: ignore
+
+# Importing the required libraries
+import io
+import streamlit as st 
+import numpy as np
+import librosa
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import Dense, Dropout
+import matplotlib.pyplot as plt
+from scipy.io.wavfile import write, read as wav_read   
+from st_audiorec import st_audiorec
+
+# Function to convert audio to spectrogram image
+def audio_to_spectrogram(file_path):
+    y, sr = librosa.load(file_path, sr=22050)
+    mel_spec = librosa.feature.melspectrogram(y=y, sr=sr, n_mels=128, hop_length=512)
+    mel_spec_db = librosa.power_to_db(mel_spec, ref=np.max)
+    plt.figure(figsize=(4, 4))
+    plt.axis('off')
+    plt.imshow(mel_spec_db, aspect='auto', origin='lower')
+    plt.tight_layout()
+    plt.savefig("spectrogram.png")
+    plt.close()
+
+# Function to create the gender classification model
+def create_model(vector_length=128):
+    model = Sequential([
+    Dense(256, input_shape=(vector_length,), activation='relu'),
+    Dropout(0.3),
+    Dense(256, activation='relu'),
+    Dropout(0.3),
+    Dense(128, activation='relu'),
+    Dropout(0.3),
+    Dense(128, activation='relu'),
+    Dropout(0.3),
+    Dense(64, activation='relu'),
+    Dropout(0.3),
+    Dense(1, activation='sigmoid')
+    ])
+    model.compile(loss='binary_crossentropy', metrics=['accuracy'], optimizer='adam')
+    model.summary()
+    return model
+
+# Load the pre-trained model
+model = create_model()
+model.load_weights("saved_model.h5")
+
+# Streamlit app
+st.title("Voice Gender Detection")
+st.write("This application detects the gender from recorded voice using a Multilayer Perceptron")
+
+# Option to upload a file
+uploaded_file = st.file_uploader("Upload an audio file", type=['wav', 'mp3'])
+
+# Function to extract features from audio file
+def extract_feature(file_name):
+    X, sample_rate = librosa.core.load(file_name)
+    result = np.array([])
+    mel = np.mean(librosa.feature.melspectrogram(y=X, sr=sample_rate).T,axis=0)
+    result = np.hstack((result, mel))
+    return result
+
+# Function to classify gender
+def classify_gender(file_path):
+    features = extract_feature(file_path).reshape(1, -1)
+    male_prob = model.predict(features, verbose=0)[0][0]
+    female_prob = 1 - male_prob
+    gender = "male" if male_prob > female_prob else "female"
+    probability = round(male_prob, 2) if gender == "male" else round(female_prob, 2)
+    return gender, probability
+
+if uploaded_file is not None:
+    with open("uploaded_audio.wav", "wb") as f:
+        f.write(uploaded_file.getbuffer())
+    st.audio(uploaded_file, format='audio/wav')
+    if st.button("Submit"):
+        audio_to_spectrogram("uploaded_audio.wav")
+        st.image("spectrogram.png", caption="Mel Spectrogram of the uploaded audio file", use_column_width="auto", width=200)
+        gender, probability = classify_gender("uploaded_audio.wav")
+        st.write(f"Predicted Gender: {gender}")
+        st.write(f"Probability: {probability}")
+
+wav_audio_data = st_audiorec()
+
+if wav_audio_data is not None:
+    # Convert byte string to numpy array
+    wav_io = io.BytesIO(wav_audio_data)
+    sr, audio_data = wav_read(wav_io)
+
+    # Save numpy array to WAV file
+    wav_file_path = "recorded_audio.wav"
+    write(wav_file_path, sr, audio_data)
+
+    st.audio(wav_audio_data, format='audio/wav')
+    audio_to_spectrogram(wav_file_path)
+    st.image("spectrogram.png", caption="Mel Spectrogram of the uploaded audio file", use_column_width="auto", width=200)
+    gender, probability = classify_gender(wav_file_path)
+    st.write(f"Predicted Gender: {gender}")
+    st.write(f"Probability: {probability}")

requirements.txt

@@ -0,0 +1,8 @@
+streamlit-audiorec
+streamlit
+librosa==0.9.2
+tensorflow
+scipy
+numpy
+matplotlib
+

saved_model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f0eadf58f6d635a4f75a4a66feae96af8939157428f606b5155511dfab6b16d
+size 1936312