Use Case and High-Level Description

A human gesture recognition model for the Jester dataset recognition scenario (gesture-level recognition). The model uses an S3D framework with MobileNet V3 backbone. Please refer to the Jester* dataset specification to see the list of gestures that are recognized by this model.

The model accepts a stack of frames (8 frames) sampled with a constant framerate (15 FPS) and produces a prediction on the input clip.

Specification

Metric	Value
Top-1 accuracy (continuous Jester*)	93.58%
GFlops	4.2269
MParams	4.1128
Source framework	PyTorch*

Input

Original Model

Batch of images of the shape [1x3x8x224x224] in the [BxCxTxHxW] format, where:

B - batch size
C - channel
T - sequence length
H - height
W - width

Channel order is RGB.

Converted Model

Batch of images of the shape [1x3x8x224x224] in the [BxCxTxHxW] format, where:

B - batch size
C - channel
T - sequence length
H - height
W - width

Channel order is RGB.

Output

The model outputs a tensor with the shape [Bx27], each row is a logits vector of performed Jester* gestures.

Original Model

Blob of the shape [1, 27] in the [BxC] format, where:

B - batch size
C - predicted logits size

Converted Model

Blob of the shape [1, 27] in the [BxC] format, where:

B - batch size
C - predicted logits size

Download a Model and Convert it into Inference Engine Format

You can download models and if necessary convert them into Inference Engine format using the Model Downloader and other automation tools as shown in the examples below.

An example of using the Model Downloader:

python3 <omz_dir>/tools/downloader/downloader.py --name <model_name>

An example of using the Model Converter:

python3 <omz_dir>/tools/downloader/converter.py --name <model_name>

Legal Information

[*] Other names and brands may be claimed as the property of others.

The original model is distributed under the Apache License 2.0.