Flexible ViG: Learning the Self-Saliency for Flexible Object Recognition

TL;DR

This paper introduces FViG, a novel neural network that enhances flexible object recognition by learning self-saliency, addressing challenges posed by diverse shapes, translucency, and ambiguous boundaries, validated on a new flexible object dataset.

Contribution

The paper proposes FViG, a self-saliency learning framework for flexible object recognition, and introduces the first Flexible Dataset for comprehensive evaluation.

Findings

FViG significantly improves flexible object recognition accuracy.

The method effectively captures shape and size variations.

Experiments on the Flexible Dataset validate the approach's effectiveness.

Abstract

Existing computer vision methods mainly focus on the recognition of rigid objects, whereas the recognition of flexible objects remains unexplored. Recognizing flexible objects poses significant challenges due to their inherently diverse shapes and sizes, translucent attributes, ambiguous boundaries, and subtle inter-class differences. In this paper, we claim that these problems primarily arise from the lack of object saliency. To this end, we propose the Flexible Vision Graph Neural Network (FViG) to optimize the self-saliency and thereby improve the discrimination of the representations for flexible objects. Specifically, on one hand, we propose to maximize the channel-aware saliency by extracting the weight of neighboring nodes, which adapts to the shape and size variations in flexible objects. On the other hand, we maximize the spatial-aware saliency based on clustering to aggregate neighborhood information for the centroid nodes, which introduces local context information for the representation learning. To verify the performance of flexible objects recognition thoroughly, for the first time we propose the Flexible Dataset (FDA), which consists of various images of flexible objects collected from real-world scenarios or online. Extensive experiments evaluated on our Flexible Dataset demonstrate the effectiveness of our method on enhancing the discrimination of flexible objects.