Meta-Learning Initializations for Image Segmentation

TL;DR

This paper extends meta-learning algorithms to image segmentation, introduces a new neural network architecture called EfficientLab, and evaluates meta-learning's effectiveness on segmentation tasks and a new benchmark dataset.

Contribution

It presents a novel meta-learning approach for image segmentation, a new architecture EfficientLab, and a benchmark dataset FP-k for evaluating meta-learning in segmentation.

Findings

Meta-learning improves few-shot segmentation performance.

EfficientLab achieves state-of-the-art results on FSS-1000.

Meta-learning performance converges to transfer learning beyond 10 examples.

Abstract

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 .