Shallow Deep Learning Can Still Excel in Fine-Grained Few-Shot Learning

TL;DR

This paper demonstrates that shallow deep networks like ConvNet-4, enhanced with a location-aware constellation module, can outperform or match deeper models in fine-grained few-shot learning tasks.

Contribution

The introduction of a location-aware constellation network with novel spatial and frequency domain encoding techniques for improved few-shot learning performance.

Findings

LCN-4 outperforms ConvNet-4 based state-of-the-art methods.

LCN-4 achieves comparable or superior results to ResNet12-based approaches.

Validation on three fine-grained benchmarks confirms effectiveness.

Abstract

Deep learning has witnessed the extensive utilization across a wide spectrum of domains, including fine-grained few-shot learning (FGFSL) which heavily depends on deep backbones. Nonetheless, shallower deep backbones such as ConvNet-4, are not commonly preferred because they're prone to extract a larger quantity of non-abstract visual attributes. In this paper, we initially re-evaluate the relationship between network depth and the ability to fully encode few-shot instances, and delve into whether shallow deep architecture could effectuate comparable or superior performance to mainstream deep backbone. Fueled by the inspiration from vanilla ConvNet-4, we introduce a location-aware constellation network (LCN-4), equipped with a cutting-edge location-aware feature clustering module. This module can proficiently encoder and integrate spatial feature fusion, feature clustering, and recessive feature location, thereby significantly minimizing the overall loss. Specifically, we innovatively put forward a general grid position encoding compensation to effectively address the issue of positional information missing during the feature extraction process of specific ordinary convolutions. Additionally, we further propose a general frequency domain location embedding technique to offset for the location loss in clustering features. We have carried out validation procedures on three representative fine-grained few-shot benchmarks. Relevant experiments have established that LCN-4 notably outperforms the ConvNet-4 based State-of-the-Arts and achieves performance that is on par with or superior to most ResNet12-based methods, confirming the correctness of our conjecture.