SOLOv2: Dynamic and Fast Instance Segmentation

TL;DR

SOLOv2 introduces a fast, dynamic instance segmentation framework that conditions mask heads on object locations and employs Matrix NMS for efficient, accurate object detection and segmentation.

Contribution

It proposes a novel dynamic mask head conditioned on object location and a new Matrix NMS method, improving speed and accuracy over previous methods.

Findings

Achieves 37.1% AP at 31.3 FPS with a lightweight model.

Outperforms state-of-the-art methods in speed and accuracy.

Demonstrates strong results in object detection and panoptic segmentation.

Abstract

In this work, we aim at building a simple, direct, and fast instance segmentation framework with strong performance. We follow the principle of the SOLO method of Wang et al. "SOLO: segmenting objects by locations". Importantly, we take one step further by dynamically learning the mask head of the object segmenter such that the mask head is conditioned on the location. Specifically, the mask branch is decoupled into a mask kernel branch and mask feature branch, which are responsible for learning the convolution kernel and the convolved features respectively. Moreover, we propose Matrix NMS (non maximum suppression) to significantly reduce the inference time overhead due to NMS of masks. Our Matrix NMS performs NMS with parallel matrix operations in one shot, and yields better results. We demonstrate a simple direct instance segmentation system, outperforming a few state-of-the-art methods in both speed and accuracy. A light-weight version of SOLOv2 executes at 31.3 FPS and yields 37.1% AP. Moreover, our state-of-the-art results in object detection (from our mask byproduct) and panoptic segmentation show the potential to serve as a new strong baseline for many instance-level recognition tasks besides instance segmentation. Code is available at: https://git.io/AdelaiDet