Skip to content

Source code for "MaskBEV: Joint Object Detection and Footprint Completion for Bird's-eye View 3D Point Clouds"

License

Notifications You must be signed in to change notification settings

norlab-ulaval/mask_bev

Repository files navigation

MaskBEV: Joint Object Detection and Footprint Completion for Bird's-eye View 3D Point Clouds

This is a work in progress migration from mmlabs libraries 1.x to 2.0.

Abstract

Recent works in object detection in LiDAR point clouds mostly focus on predicting bounding boxes around objects. This prediction is commonly achieved using anchor-based or anchor-free detectors that predict bounding boxes, requiring significant explicit prior knowledge about the objects to work properly. To remedy these limitations, we propose MaskBEV, a bird's-eye view (BEV) mask-based object detector neural architecture. MaskBEV predicts a set of BEV instance masks that represent the footprints of detected objects. Moreover, our approach allows object detection and footprint completion in a single pass. MaskBEV also reformulates the detection problem purely in terms of classification, doing away with regression usually done to predict bounding boxes. We evaluate the performance of MaskBEV on both SemanticKITTI and KITTI datasets while analyzing the architecture advantages and limitations.

Documentation

Follow dataset installation instructions to download and prepare the datasets.

Follow installation instructions to install the dependencies.

Follow training instructions to start training and evaluating MaskBEV.

Follow configuration instructions to understand the configuration options.

Follow testing instructions to test the trained models.

About

Source code for "MaskBEV: Joint Object Detection and Footprint Completion for Bird's-eye View 3D Point Clouds"

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages