Few-Shot Segmentation Without Meta-Learning: A Good Transductive Inference Is All You Need?

By Malik Boudiaf, Hoel Kervadec, Imtiaz Masud Ziko, Pablo Piantanida, Ismail Ben Ayed, Jose Dolz. 2020

240 citations recorded by the source corpus.

We show that the way inference is performed in few-shot segmentation tasks has a substantial effect on performances—an aspect often overlooked in the literature in favor of the meta-learning paradigm. We introduce a transductive inference for a given query image, leveraging the statistics of its unlabeled pixels, by optimizing a new loss containing three complementary terms: i) the cross-entropy on the labeled support pixels; ii) the Shannon entropy of the posteriors on the unlabeled query-image pixels; and iii) a global KL-divergence regularizer based on the proportion of the predicted foreground. As our inference uses a simple linear classifier of the extracted features, its computational load is comparable to inductive inference and can be used on top of any base training. Foregoing episodic training and using only standard cross-entropy training on the base classes, our inference yields competitive performances on standard benchmarks in the 1-shot scenarios. As the number of available shots increases, the gap in performances widens: on PASCAL-5i, our method brings about 5% and 6% improvements over the state-of-the-art, in the 5- and 10-shot scenarios, respectively. Furthermore, we introduce a new setting that includes domain shifts, where the base and novel classes are drawn from different datasets. Our method achieves the best performances in this more realistic setting. Our code is freely available online: https://github.com/mboudiaf/RePRI-for-Few-Shot-Segmentation.

Subfields: Computer Vision and Pattern Recognition, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Meta & Continual Learning

Citations in the corpus

Language-driven Semantic Segmentation (2022, background): “…et al., 2017; Rakelly et al., 2018; Siam et al., 2019; Wang et al., 2019; Zhang et al., 2019; Nguyen & Todorovic, 2019; Liu et al., 2020b; Wang et al., 2020; Tian et al., 2020; Boudiaf et al., 2021; Min et al., 2021) offer ways to learn to segment novel classes based on only a few labeled images.”
Hypercorrelation Squeeze for Few-Shot Segmenation (2021, methodology): “8%p of mIoU improvements over [70] and [4], verifying its superiority in few-shot segmentation task.”
Generalized Few-shot Semantic Segmentation (2020, background): “It is worth noting that some recently proposed advanced frameworks, e.g ., RePRI [3] and HSNet [23], are not practical to be applied to the setting of GFS-Seg because their inference processes require independent reasoning with the full support feature maps, instead of the prototypical feature vecto”