Update README.md

Author: vishalshar

README.md

@@ -21,13 +21,42 @@ Audio dataset given has very high frame rate, on an average every file had 80,00
 ![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/audio_preprocessing-1.png)
 
 
+
+
+### Using ANN
+
+During initial experiments ANN was not performing good and later after several experiments a Multi Layer Perceptron (MLP) model was build based on intuition of CNN. Before we feed audio data in network it was max pooled in 3 different layers and output of pooled layers was given input to the fully connected layers as shows in below figure. To merge features extracted from different pooling layers output of fully connected layer was merged.
+
+
+#### Core Idea: 
+Sample Bee Audio and expected feature extraction using pooling layers and merging fully connected layers
+
+![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/audio_graph-1.png)
+
+
+#### Performance: 
+
+Training was done for 500 epochs using Adaptive Moment Estimation (adam) as optimizer with learning rate of 0.0005. Figure 9 displays accuracy during training.
+
+|            | Training           | Testing        |
+|-----| ------------- |:-------------:|
+|Accuracy   | 91.11%      | 88.25%       |
+
+![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/ANN_Net_2-1.png)
+
+#### Accuracy
+
+![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/bee_ann_audio.png)
+
+
+
 ### Using CNN
 
 A network using Convolution layers was used to build classifier, network architecture is shown in Fig 6. The number of filters for both convolution was 64 and filter_size was 10 and 3 for respective layers followed by 3 fully connected layers, details about activation function used is in code. Max pooling was used after each convolution layer. During training over fitting was observed, to handle that dropout of 50% (keep) was used after first two fully connected layers and also ‘L2’ regularization was added to both layers. Input length was fixed as 22,000 with 1 channel. During training it was also observed, without downsampling
 data model was not able to generalize well between bee and noise data. Adding downsampling technique helped the
 model in generalization.
 
-## Performance: 
+#### Performance: 
 
 Training was done for 500 epochs using Adaptive Moment Estimation (adam) as optimizer with learning rate of 0.0001. 
 
@@ -37,9 +66,8 @@ Training was done for 500 epochs using Adaptive Moment Estimation (adam) as opti
 |Accuracy   | 99.88%      | 99.45%       |
 
 
-
-![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/ANN_Net_2-1.png)
 ![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/CNN_Net-1.png)
-![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/audio_graph-1.png)
-![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/bee_ann_audio.png)
+
+#### Accuracy
+
 ![alt text](https://raw.githubusercontent.com/vishalshar/Audio-Classification-using-CNN-MLP/master/img/bee_cnn_audio.png)