# FSS-1000: A 1000-Class Dataset for Few-Shot Segmentation

## TL;DR

This paper introduces FSS-1000, a new dataset with 1000 object classes for few-shot segmentation, demonstrating that training from scratch on this dataset can outperform pre-trained models on limited data.

## Contribution

The creation of FSS-1000, a large-scale dataset for few-shot segmentation with diverse objects, and showing that models trained from scratch can outperform pre-trained models on this dataset.

## Key findings

- Training from scratch on FSS-1000 yields comparable or better results than using ImageNet pre-trained weights.
- FSS-1000 includes many objects not present in previous datasets, such as tiny objects and logos.
- Simple models trained on FSS-1000 are effective and easily extendable for new classes.

## Abstract

Over the past few years, we have witnessed the success of deep learning in image recognition thanks to the availability of large-scale human-annotated datasets such as PASCAL VOC, ImageNet, and COCO. Although these datasets have covered a wide range of object categories, there are still a significant number of objects that are not included. Can we perform the same task without a lot of human annotations? In this paper, we are interested in few-shot object segmentation where the number of annotated training examples are limited to 5 only. To evaluate and validate the performance of our approach, we have built a few-shot segmentation dataset, FSS-1000, which consists of 1000 object classes with pixelwise annotation of ground-truth segmentation. Unique in FSS-1000, our dataset contains significant number of objects that have never been seen or annotated in previous datasets, such as tiny daily objects, merchandise, cartoon characters, logos, etc. We build our baseline model using standard backbone networks such as VGG-16, ResNet-101, and Inception. To our surprise, we found that training our model from scratch using FSS-1000 achieves comparable and even better results than training with weights pre-trained by ImageNet which is more than 100 times larger than FSS-1000. Both our approach and dataset are simple, effective, and easily extensible to learn segmentation of new object classes given very few annotated training examples. Dataset is available at https://github.com/HKUSTCV/FSS-1000.