P3DC-Shot: Prior-Driven Discrete Data Calibration for Nearest-Neighbor Few-Shot Classification

TL;DR

P3DC-Shot introduces a prior-driven data calibration technique for nearest-neighbor few-shot classification, improving robustness by calibrating support data based on base class prototypes, achieving competitive results without extra learning.

Contribution

The paper proposes a novel discrete data calibration method leveraging base class prototypes to enhance NN-based few-shot classification without additional training.

Findings

Outperforms or matches state-of-the-art methods on various datasets.

Effective calibration improves classification accuracy in few-shot scenarios.

No additional learning steps required, reducing complexity.

Abstract

Nearest-Neighbor (NN) classification has been proven as a simple and effective approach for few-shot learning. The query data can be classified efficiently by finding the nearest support class based on features extracted by pretrained deep models. However, NN-based methods are sensitive to the data distribution and may produce false prediction if the samples in the support set happen to lie around the distribution boundary of different classes. To solve this issue, we present P3DC-Shot, an improved nearest-neighbor based few-shot classification method empowered by prior-driven data calibration. Inspired by the distribution calibration technique which utilizes the distribution or statistics of the base classes to calibrate the data for few-shot tasks, we propose a novel discrete data calibration operation which is more suitable for NN-based few-shot classification. Specifically, we treat the prototypes representing each base class as priors and calibrate each support data based on its similarity to different base prototypes. Then, we perform NN classification using these discretely calibrated support data. Results from extensive experiments on various datasets show our efficient non-learning based method can outperform or at least comparable to SOTA methods which need additional learning steps.