Image Embedding
Image Embedding with Deep Neural Network Architectures
Steps through run the code
- PATH is the path where this .ipynb should be placed. In this code PATH = '/content/drive/MyDrive/Colab Notebooks'
- There should be folder named '/embedding-logs' in the PATH folder that would contain the log files for embedding.
- Xception Model [1] has been used to extract the features and classification.
- Bangla Handwritten Nurerals Dataset has been used in this example. But any other image dataset of which the labels are separated by different folders can be visualized with this script.
- The image dimension is kept (3, 28, 28).
- After run the TensorBoard, select "Projector" option to visualize the embedding.
3D Visualization of Embedding
2D Visualization of Embedding
Reference
[1] Chollet, François. "Xception: Deep learning with depthwise separable convolutions." Proceedings of the IEEE conference on computer vision and pattern recognition. 2017.