README.rst

Instance Segmentation Pipeline

Overview

instance_segmentation.sh demonstrates instance segmentation on one video file source and verifies Hailo’s configuration.

This is done by running a single-stream instance segmentation pipeline on top of GStreamer using the Hailo-8 device.

Options

./instance_segmentation.sh [--input FILL-ME]

• --input is an optional flag, a path to the video displayed (default is detection.mp4).
• --show-fps is an optional flag that enables printing FPS on screen.
• --print-gst-launch is a flag that prints the ready gst-launch command without running it"

Run

cd $TAPPAS_WORKSPACE/apps/gstreamer/general/instance_segmentation
./instance_segmentation.sh

The output should look like:

Model

• yolact_regnetx_800mf_20classes - https://github.com/hailo-ai/hailo_model_zoo/blob/master/hailo_model_zoo/cfg/networks/yolact_regnetx_800mf_20classes.yaml

How does it work?

This app is based on our single network pipeline template

How to use Retraining to replace models

Note

It is recommended to first read the Retraining TAPPAS Models page.

You can use Retraining Dockers (available on Hailo Model Zoo), to replace the following models with ones that are trained on your own dataset:

• yolact_regnetx_800mf

  • Retraining docker

  • For best compatibility and performance with TAPPAS, use for compilation the corresponsing YAML file from above.

  • TAPPAS changes to replace model:
    • Update HEF_PATH on the .sh file
    • Update yolact.cpp with your new paremeters, then recompile to create libyolact_post.so