Fast and Precise Binary Instance Segmentation of 2D Objects for Automotive Applications

TL;DR

This paper presents a real-time CPU-based binary instance segmentation method for 2D objects in automotive applications, utilizing extreme points to improve accuracy and speed over existing encoder-decoder networks.

Contribution

It introduces a novel approach that uses extreme points as input to enhance segmentation quality and efficiency on CPUs for automotive tasks.

Findings

Achieves higher IoU than existing encoder-decoder methods.

Runs in real-time on CPU hardware.

Simplifies labeling by using extreme points instead of bounding boxes.

Abstract

In this paper, we focus on improving binary 2D instance segmentation to assist humans in labeling ground truth datasets with polygons. Humans labeler just have to draw boxes around objects, and polygons are generated automatically. To be useful, our system has to run on CPUs in real-time. The most usual approach for binary instance segmentation involves encoder-decoder networks. This report evaluates state-of-the-art encoder-decoder networks and proposes a method for improving instance segmentation quality using these networks. Alongside network architecture improvements, our proposed method relies upon providing extra information to the network input, so-called extreme points, i.e. the outermost points on the object silhouette. The user can label them instead of a bounding box almost as quickly. The bounding box can be deduced from the extreme points as well. This method produces better IoU compared to other state-of-the-art encoder-decoder networks and also runs fast enough when it is deployed on a CPU.