Discwise Active Learning for LiDAR Semantic Segmentation

TL;DR

This paper introduces Discwise Active Learning (DiAL) for LiDAR semantic segmentation, optimizing annotation efficiency by selecting regions across frames and addressing challenges like point density variation and frame selection.

Contribution

The paper presents a novel discwise active learning framework with a new acquisition function, a mixed-integer linear program for frame selection, and a semi-supervised learning approach.

Findings

Effective region selection improves annotation efficiency.

The acquisition function accounts for point density variations.

Semi-supervised learning enhances segmentation performance.

Abstract

While LiDAR data acquisition is easy, labeling for semantic segmentation remains highly time consuming and must therefore be done selectively. Active learning (AL) provides a solution that can iteratively and intelligently label a dataset while retaining high performance and a low budget. In this work we explore AL for LiDAR semantic segmentation. As a human expert is a component of the pipeline, a practical framework must consider common labeling techniques such as sequential labeling that drastically improve annotation times. We therefore propose a discwise approach (DiAL), where in each iteration, we query the region a single frame covers on global coordinates, labeling all frames simultaneously. We then tackle the two major challenges that emerge with discwise AL. Firstly we devise a new acquisition function that takes 3D point density changes into consideration which arise due to location changes or ego-vehicle motion. Next we solve a mixed-integer linear program that provides a general solution to the selection of multiple frames while taking into consideration the possibilities of disc intersections. Finally we propose a semi-supervised learning approach to utilize all frames within our dataset and improve performance.