TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK	REMOVE
Few-Shot Semantic Segmentation	FSS-1000	EfficientLab	Mean IoU	82.78	# 3

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/meta-learning-initializations-for-image-1/few-shot-semantic-segmentation-on-fss-1000)](https://paperswithcode.com/sota/few-shot-semantic-segmentation-on-fss-1000?p=meta-learning-initializations-for-image-1)`

Meta-Learning Initializations for Image Segmentation

13 Dec 2019 · Sean M. Hendryx, Andrew B. Leach, Paul D. Hein, Clayton T. Morrison ·

We extend first-order model agnostic meta-learning algorithms (including FOMAML and Reptile) to image segmentation, present a novel neural network architecture built for fast learning which we call EfficientLab, and leverage a formal definition of the test error of meta-learning algorithms to decrease error on out of distribution tasks. We show state of the art results on the FSS-1000 dataset by meta-training EfficientLab with FOMAML and using Bayesian optimization to infer the optimal test-time adaptation routine hyperparameters. We also construct a small benchmark dataset, FP-k, for the empirical study of how meta-learning systems perform in both few- and many-shot settings. On the FP-k dataset, we show that meta-learned initializations provide value for canonical few-shot image segmentation but their performance is quickly matched by conventional transfer learning with performance being equal beyond 10 labeled examples. Our code, meta-learned model, and the FP-k dataset are available at https://github.com/ml4ai/mliis .

PDF Abstract