Emotion Recognition

Our new model ZOO works with DepthAI V3. Find out more in our documentation.

0 Likes

Model Details

Model Description

Emotion-Recogniton-8-ENet model is based on the EfficientNet convolutional neural network model. It can recognize 8 emotions expressed with face gestures. 8 emotions include anger, contempt, disgust, fear, happiness, neutral, sadness, and surprise.

Developed by: A. Savchenko
Shared by:
Model type: Computer vision
License:
Resources for more information:

Training Details

Training Data

The model was trained on dataset. The dataset contains 3.31 million images of 9131 subjects (identities), with an average of 362.6 images for each subject.

Testing Details

Metrics

The evaluation is performed on

Metric	Value
Accuracy (8 classes)	63.13%
Accuracy (7 classes)	66.51%

For more information please check the .

Technical Specifications

Input/Output Details

Input:
- Name: input
  - Info: NCHW BGR 0-255 image.
Output:
- Name: output
  - Info: Prediction scores for every emotion.

Model Architecture

The general Efficient-net network consists of:

Stem: Initial layer with a standard convolution followed by a batch normalization and a ReLU6 activation.
Body: Consists of a series of MBConv blocks with different configurations. Each block includes depthwise separable convolutions and squeeze-and-excitation layers.
Head: Includes a final convolutional block, followed by a global average pooling layer.

For specifics about modified model check .

Throughput

Model variant: emotion-recognition:gray-64x64

• Input shape: [1, 1, 64, 64] • Output shape: [1, 8]

• Params (M): 8.758 • GFLOPs: 0.878

Platform	Precision	Throughput (infs/sec)	Power Consumption (W)
RVC2	FP16	220.04	N/A

Model variant: emotion-recognition:260x260

• Input shape: ['batch_size', 3, 260, 260] • Output shape: ['batch_size', 8]

Platform	Precision	Throughput (infs/sec)	Power Consumption (W)
RVC2	FP16	11.41	N/A
RVC4	FP32	271.35	2.41

* Benchmarked with , using 2 threads (and the DSP runtime in balanced mode for RVC4).

* Parameters and FLOPs are obtained from the package.

Quantization

RVC4 version of the model was quantized using a custom dataset. This was created by taking 40 face-cropped images from different datasets available on the web.

Utilization

Models converted for RVC Platforms can be used for inference on OAK devices. DepthAI pipelines are used to define the information flow linking the device, inference model, and the output parser (as defined in model head(s)). Below, we present the most crucial utilization steps for the particular model. Please consult the docs for more information.

Install DAIv3 and depthai-nodes libraries:

pip install depthai
pip install depthai-nodes

Define model:

model_description = dai.NNModelDescription(
    "luxonis/emotion-recognition:260x260"
)

nn = pipeline.create(ParsingNeuralNetwork).build(
    <CameraNode>, model_description
)

Inspect model head(s):

ClassificationParser that outputs message (predicted emotion classes and scores.).

Get parsed output(s):

while pipeline.isRuning():
    parser_output: Classifications = parser_output_queue.get()

Example

You can quickly run the model using our example.

The example demonstrates how to build a 2-stage DepthAI pipeline consisting of a face detection model and an emotion recognition model. It automatically downloads the models, creates a DepthAI pipeline, runs the inference, and displays the results using our DepthAI visualizer tool.

To try it out, run:

python3 main.py

Emotion recognition model.
License	Apache 2.0 Commercial use
Downloads	509
Tasks	Classification
Model Types	ONNX

Model Variants

Name	Version	Available For	Created At	Deploy
		RVC2	10 months ago
		RVC2, RVC4	10 months ago