lidar_apollo_segmentation_tvm

Design

Usage

Neural network

This package will not build without a neural network for its inference. The network is provided by the cmake function exported by the tvm_utility package. See its design page for more information on how to enable downloading pre-compiled networks (by setting the DOWNLOAD_ARTIFACTS cmake variable), or how to handle user-compiled networks.

Backend

The backend used for the inference can be selected by setting the lidar_apollo_segmentation_tvm_BACKEND cmake variable. The current available options are llvm for a CPU backend, and vulkan for a GPU backend. It defaults to llvm.

Convolutional Neural Networks (CNN) Segmentation

See the original design by Apollo. The paragraph of interest goes up to, but excluding, the "MinBox Builder" paragraph. This package instead relies on further processing by a dedicated shape estimator.

Note: the parameters described in the original design have been modified and are out of date.

Inputs / Outputs / API

The package exports a boolean lidar_apollo_segmentation_tvm_BUILT cmake variable.

Reference

Lidar segmentation is based off a core algorithm by Apollo, with modifications from [TIER IV] (https://github.com/tier4/lidar_instance_segmentation_tvm) for the TVM backend.