Associative Embedding: End-to-End Learning for Joint Detection and Grouping

TL;DR

Associative embedding is a new end-to-end learning method that enables neural networks to simultaneously detect and group objects, improving performance in tasks like pose estimation and segmentation.

Contribution

The paper introduces associative embedding, a novel approach for joint detection and grouping that integrates seamlessly into existing neural network architectures.

Findings

Achieves state-of-the-art multi-person pose estimation results on MPII and MS-COCO datasets.

Simplifies the detection and grouping pipeline by eliminating multi-stage processes.

Demonstrates versatility across multiple computer vision tasks.

Abstract

We introduce associative embedding, a novel method for supervising convolutional neural networks for the task of detection and grouping. A number of computer vision problems can be framed in this manner including multi-person pose estimation, instance segmentation, and multi-object tracking. Usually the grouping of detections is achieved with multi-stage pipelines, instead we propose an approach that teaches a network to simultaneously output detections and group assignments. This technique can be easily integrated into any state-of-the-art network architecture that produces pixel-wise predictions. We show how to apply this method to both multi-person pose estimation and instance segmentation and report state-of-the-art performance for multi-person pose on the MPII and MS-COCO datasets.