fastseg-small#

Use Case and High-Level Description#

fastseg-small is an accurate real-time semantic segmentation model, pre-trained on Cityscapes dataset for 19 object classes, listed in <omz_dir>/data/dataset_classes/cityscapes_19cl_bkgr.txt file. See Cityscapes classes definition for more details. The model was built on MobileNetV3 small backbone and modified segmentation head based on LR-ASPP. This model can be used for efficient segmentation on a variety of real-world street images. For model implementation details see original repository.

Specification#

Metric

Value

Type

Semantic segmentation

GOps

69.2204

MParams

1.1

Source framework

PyTorch*

Accuracy#

Metric

Value

mean_iou

67.15%

Input#

Original model#

Image, name: input0, shape: 1, 3, 1024, 2048, format: B, C, H, W, where:

  • B - batch size

  • C - number of channels

  • H - image height

  • W - image width

Expected color order: RGB. Mean values: [123.675, 116.28, 103.53], scale values: [58.395, 57.12, 57.375]

Converted Model#

Image, name: input0, shape: 1, 3, 1024, 2048, format: B, C, H, W, where:

  • B - batch size

  • C - number of channels

  • H - image height

  • W - image width

Expected color order: BGR.

Output#

Original Model#

Float values, which represent scores of a predicted class for each image pixel. The model was trained on Cityscapes dataset with 19 categories of objects. Name: output0, shape: 1, 19, 1024, 2048 in B, N, H, W format, where:

  • B - batch size

  • N - number of classes

  • H - image height

  • W - image width

Converted Model#

Float values, which represent scores of a predicted class for each image pixel. The model was trained on Cityscapes dataset with 19 categories of objects. Name: output0, shape: 1, 19, 1024, 2048 in B, N, H, W format, where:

  • B - batch size

  • N - number of classes

  • H - image height

  • W - image width

Download a Model and Convert it into OpenVINO™ IR Format#

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

An example of using the Model Downloader:

omz_downloader --name <model_name>

An example of using the Model Converter:

omz_converter --name <model_name>

Demo usage#

The model can be used in the following demos provided by the Open Model Zoo to show its capabilities: