# LVIS: A Dataset for Large Vocabulary Instance Segmentation

## TL;DR

LVIS introduces a large-scale dataset with over 2 million instance segmentation masks across 1000+ categories, emphasizing the challenge of recognizing rare categories with few samples in natural images.

## Contribution

The paper presents LVIS, a new dataset for large vocabulary instance segmentation, highlighting its long-tail distribution and the challenge it poses for current deep learning methods.

## Key findings

- LVIS contains over 2 million masks for 1000+ categories.
- The dataset exhibits a Zipfian distribution with many categories having few samples.
- It provides a new benchmark for low-sample recognition in complex scenes.

## Abstract

Progress on object detection is enabled by datasets that focus the research community's attention on open challenges. This process led us from simple images to complex scenes and from bounding boxes to segmentation masks. In this work, we introduce LVIS (pronounced `el-vis'): a new dataset for Large Vocabulary Instance Segmentation. We plan to collect ~2 million high-quality instance segmentation masks for over 1000 entry-level object categories in 164k images. Due to the Zipfian distribution of categories in natural images, LVIS naturally has a long tail of categories with few training samples. Given that state-of-the-art deep learning methods for object detection perform poorly in the low-sample regime, we believe that our dataset poses an important and exciting new scientific challenge. LVIS is available at http://www.lvisdataset.org.