Few-shot Learning for Domain-specific Fine-grained Image Classification

TL;DR

This paper introduces a novel feature fusion model with focus mechanisms and a specialized loss function for few-shot, fine-grained image classification, validated on a real-world marine ecological dataset.

Contribution

The paper proposes a new model combining focus area localization, high-order feature integration, and a Center Neighbor Loss for improved few-shot fine-grained classification.

Findings

Achieves competitive performance on the miniPPlankton dataset.

Demonstrates effectiveness of focus and high-order features in fine-grained tasks.

Validates approach with extensive experiments.

Abstract

Learning to recognize novel visual categories from a few examples is a challenging task for machines in real-world industrial applications. In contrast, humans have the ability to discriminate even similar objects with little supervision. This paper attempts to address the few shot fine-grained image classification problem. We propose a feature fusion model to explore discriminative features by focusing on key regions. The model utilizes the focus area location mechanism to discover the perceptually similar regions among objects. High-order integration is employed to capture the interaction information among intra-parts. We also design a Center Neighbor Loss to form robust embedding space distributions. Furthermore, we build a typical fine-grained and few-shot learning dataset miniPPlankton from the real-world application in the area of marine ecological environments. Extensive experiments are carried out to validate the performance of our method. The results demonstrate that our model achieves competitive performance compared with state-of-the-art models. Our work is a valuable complement to the model domain-specific industrial applications.