reab5555 commited on
Commit
9f4919a
·
verified ·
1 Parent(s): 2bfe741

Update visualization.py

Browse files
Files changed (1) hide show
  1. visualization.py +0 -34
visualization.py CHANGED
@@ -154,40 +154,6 @@ def plot_combined_mse(df, mse_embeddings, mse_posture, mse_voice, title):
154
  plt.close()
155
  return fig
156
 
157
- def combined_heatmap(mse_embeddings, mse_posture, mse_voice, df):
158
- # Normalize MSE values
159
- mse_embeddings_norm = (mse_embeddings - np.min(mse_embeddings)) / (np.max(mse_embeddings) - np.min(mse_embeddings))
160
- mse_posture_norm = (mse_posture - np.min(mse_posture)) / (np.max(mse_posture) - np.min(mse_posture))
161
- mse_voice_norm = (mse_voice - np.min(mse_voice)) / (np.max(mse_voice) - np.min(mse_voice))
162
-
163
- combined_mse = np.zeros((3, len(mse_embeddings)))
164
- combined_mse[0] = mse_embeddings_norm
165
- combined_mse[1] = mse_posture_norm
166
- combined_mse[2] = mse_voice_norm
167
-
168
- plt.figure(figsize=(20, 3), dpi=300)
169
- fig, ax = plt.subplots(figsize=(20, 3))
170
-
171
- # Create heatmap
172
- sns.heatmap(combined_mse, cmap='Reds', cbar=False, ax=ax)
173
-
174
- # Set y-axis labels
175
- ax.set_yticks([0.5, 1.5, 2.5])
176
- ax.set_yticklabels(['Facial Features', 'Body Posture', 'Voice'], fontsize=8)
177
-
178
- # Set x-axis ticks to timecodes
179
- num_ticks = min(60, len(mse_embeddings))
180
- tick_locations = np.linspace(0, len(mse_embeddings) - 1, num_ticks).astype(int)
181
- tick_labels = [df['Timecode'].iloc[i] if i < len(df) else '' for i in tick_locations]
182
-
183
- ax.set_xticks(tick_locations)
184
- ax.set_xticklabels(tick_labels, rotation=90, ha='center', va='top')
185
-
186
- ax.set_title('Combined MSE Heatmap')
187
-
188
- plt.tight_layout()
189
- plt.close(fig)
190
- return fig
191
 
192
  def plot_mse_histogram(mse_values, title, anomaly_threshold, color='blue'):
193
  plt.figure(figsize=(16, 3), dpi=300)
 
154
  plt.close()
155
  return fig
156
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
157
 
158
  def plot_mse_histogram(mse_values, title, anomaly_threshold, color='blue'):
159
  plt.figure(figsize=(16, 3), dpi=300)