Skip to content

Latest commit

 

History

History
 
 

emotion_ferplus

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 
model.onnx
model.onnx
 
 
test_data_0.npz
test_data_0.npz
 
 
test_data_1.npz
test_data_1.npz
 
 
test_data_2.npz
test_data_2.npz
 
 

README.md

FER+ Emotion Recognition

Download: https://www.cntk.ai/OnnxModels/emotion_ferplus.tar.gz

Description

This model is a deep convolutional neural network for emotion recognition in faces. It is trained on the FER+ annotations for the standard Emotion FER dataset, as described in this paper.

Source

The model is trained in CNTK. You can find the source code here.

Model input and output

The model expects a grayscale input image of the shape (1x64x64), normalized to pixel values between [-1, 1]. To normalize the input image, the following computation is performed: (image - 127.5)/127.5.

Sets of sample input and output files are provided in .npz format (test_data_*.npz). The input is a normalized (1x64x64) numpy array of a test image, while the output is an array of length 8 corresponding to the output of evaluating the model on the sample input.

License

MIT