GRM: Gradient Rectification Module for Visual Place Retrieval

TL;DR

This paper introduces the Gradient Rectification Module (GRM) to improve visual place retrieval by encouraging more uniform descriptor distribution, addressing the issue of low-dimensional principal space caused by gradient degradation.

Contribution

The paper proposes GRM, a novel module that rectifies gradients to enhance descriptor diversity and retrieval accuracy in visual place retrieval tasks.

Findings

GRM improves retrieval performance across multiple datasets.

GRM promotes more uniform descriptor distribution.

Method generalizes to classification tasks under prototype learning.

Abstract

Visual place retrieval aims to search images in the database that depict similar places as the query image. However, global descriptors encoded by the network usually fall into a low dimensional principal space, which is harmful to the retrieval performance. We first analyze the cause of this phenomenon, pointing out that it is due to degraded distribution of the gradients of descriptors. Then, we propose Gradient Rectification Module(GRM) to alleviate this issue. GRM is appended after the final pooling layer and can rectify gradients to the complementary space of the principal space. With GRM, the network is encouraged to generate descriptors more uniformly in the whole space. At last, we conduct experiments on multiple datasets and generalize our method to classification task under prototype learning framework.