Meta-tuning Loss Functions and Data Augmentation for Few-shot Object Detection

TL;DR

This paper introduces a meta-learning approach to optimize loss functions and data augmentations for few-shot object detection, enhancing fine-tuning methods with improved performance and interpretability.

Contribution

It proposes a novel training scheme that tunes loss functions and augmentations via meta-learning, boosting few-shot detection while maintaining simplicity and interpretability.

Findings

Significant improvements on Pascal VOC and MS-COCO datasets.

Enhanced standard and generalized few-shot detection performance.

Better interpretability of loss functions compared to complex meta-models.

Abstract

Few-shot object detection, the problem of modelling novel object detection categories with few training instances, is an emerging topic in the area of few-shot learning and object detection. Contemporary techniques can be divided into two groups: fine-tuning based and meta-learning based approaches. While meta-learning approaches aim to learn dedicated meta-models for mapping samples to novel class models, fine-tuning approaches tackle few-shot detection in a simpler manner, by adapting the detection model to novel classes through gradient based optimization. Despite their simplicity, fine-tuning based approaches typically yield competitive detection results. Based on this observation, we focus on the role of loss functions and augmentations as the force driving the fine-tuning process, and propose to tune their dynamics through meta-learning principles. The proposed training scheme, therefore, allows learning inductive biases that can boost few-shot detection, while keeping the advantages of fine-tuning based approaches. In addition, the proposed approach yields interpretable loss functions, as opposed to highly parametric and complex few-shot meta-models. The experimental results highlight the merits of the proposed scheme, with significant improvements over the strong fine-tuning based few-shot detection baselines on benchmark Pascal VOC and MS-COCO datasets, in terms of both standard and generalized few-shot performance metrics.