Context Information Refinement for Few-Shot Object Detection in Remote Sensing Images

Wang, Yan; Xu, Chaofei; Liu, Cuiwei; Li, Zhaokui

doi:10.3390/rs14143255

Cited by 30 publications

(18 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, we evaluated the developed approach in combination with baseline approaches using the DIOR dataset and the NWPU VHR-10 dataset in 3-shot, 5-shot, and 10-shot cases, as illustrated in Table II and Table III. Notably, the results denoted with an asterisk (*) were originally reported in Wang's [39] paper; the results without an asterisk were obtained through rigorous testing in our experimental environment. This comparative analysis enables a robust assessment of the efficiency and performance of our suggested approach.…”

Section: Comparison With Baseline Methodsmentioning

confidence: 91%

FSOD4RSI: Few-Shot Object Detection for Remote Sensing Images via Features Aggregation and Scale Attention

Gao,

Wu,

Wang

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Due to the continuous development of few-shot learning, there have been notable advancements in methods for few-shot object detection in recent years. However, most existing methods in this domain primarily focus on natural images, neglecting the challenges posed by variations in object scales, which are usually encountered in remote sensing images. This paper proposes a new few-shot object detection model designed to handle the issue of object scale variation in remote sensing images. Our developed model has two essential parts: a feature aggregation module (FAM) and a scale-aware attention module (SAM). Considering the few-shot features of remote sensing images, we designed the FAM to improve the support and query features through channel multiplication operations utilizing a feature pyramid network (FPN) and a Transformer encoder. The created FAM better extracts the global features of remote sensing images and enhances the significant feature representation of fewshot remote sensing objects. Additionally, we design the SAM to address the scale variation problems that frequently occur in remote sensing images. By employing multiscale convolutions, the SAM enables the acquisition of contextual features while adapting to objects of varying scales. Extensive experiments were conducted on benchmark datasets, including NWPU VHR-10 and DIOR datasets, and the results show that our model indeed addresses the challenges posed by object scale variation and improves the applicability of few-shot object detection in the remote sensing domain.

show abstract

Section: Comparison With Baseline Methodsmentioning

confidence: 91%

FSOD4RSI: Few-Shot Object Detection for Remote Sensing Images via Features Aggregation and Scale Attention

Gao,

Wu,

Wang

et al. 2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…To demonstrate the superior performance of the proposed GC-IKDH model significantly, we compared it with state-of-the-art few-shot object detection models in RSI, including FSODM [56], PAMS-Det [57], OAF [58], CIR-FSD [59], SAM-BFS [54], and TSF-RGR [60]. Additionally, we compared it with the original Fast R-CNN [23] without fine-tuning.…”

Section: Comparison With Sota Experimentsmentioning

confidence: 99%

Few‐shot object detection based on global context and implicit knowledge decoupled head

Li,

Yang,

Liu

et al. 2024

IET Image Processing

View full text Add to dashboard Cite

The acquisition cycle of remote sensing images is slow, and the labelling process encounters challenges, which have become prominent with the rapid development of remote sensing image object detection research. Therefore, this article provides a way to make the model better capture the diversity and contextual relationships in the data, and solve this problem by more than just data augmentation. Specifically, this method is a few‐shot object detection method for remote sensing based on global context combined with implicit knowledge decoupled head (GC‐IKDH). This method first uses a segmentation strategy to convert high‐resolution images into low‐resolution images and expands the sample size through a generative model. Secondly, GC attention is introduced to generate a GC vector by weighting and averaging the information of each position in the input sequence, which helps the model better understand the semantics of the input sequence. Finally, an IKDH is added to improve the model head, which is used to learn specific features in the data so that the model can better handle the diversity in the data. Experimental results show that GC attention and IKDH boosting provide a good performance boost to the baseline model. Compared with other few‐shot samples, this method achieves state‐of‐the‐art performance under different shot settings and highly competitive results on two benchmark datasets (NWPU VHR‐10 and DIOR).

show abstract

“…To address this limitation, people have been exploring alternative techniques that require fewer RSI annotations. These techniques include semisupervised learning [9,10,19], weakly-supervised learning [20,21,22,23,24], few-shot learning [25,26,27,28,29,30] and active learning [31,32,33],. These methods aim to minimize annotation expenses and speed up the labeling process, making them highly valuable for practical applications in remote sensing.…”

Section: Related Work a Label Efficient Object Detection In Rsismentioning

confidence: 99%

Feature Alignment FPN for Oriented Object Detection in Remote Sensing Images

et al. 2023

IEEE Geosci. Remote Sensing Lett.

View full text Add to dashboard Cite

Annotating remote sensing images (RSIs) presents a notable challenge due to its labor-intensive nature. Semisupervised object detection (SSOD) methods tackle this issue by generating pseudo-labels for the unlabeled data, assuming that all classes found in the unlabeled dataset are also represented in the labeled data. However, real-world situations introduce the possibility of out-of-distribution (OOD) samples being mixed with in-distribution (ID) samples within the unlabeled dataset. In this paper, we delve into techniques for conducting SSOD directly on uncurated unlabeled data, which is termed Open-Set Semi-Supervised Object Detection (OSSOD). Our approach commences by employing labeled in-distribution data to dynamically construct a class-wise feature bank (CFB) that captures features specific to each class. Subsequently, we compare the features of predicted object bounding boxes with the corresponding entries in the CFB to calculate OOD scores. We design an adaptive threshold based on the statistical properties of the CFB, allowing us to filter out OOD samples effectively. The effectiveness of our proposed method is substantiated through extensive experiments on two widely used remote sensing object detection datasets: DIOR and DOTA. These experiments showcase the superior performance and efficacy of our approach for OSSOD on RSIs.

show abstract

Context Information Refinement for Few-Shot Object Detection in Remote Sensing Images

Cited by 30 publications

References 47 publications

FSOD4RSI: Few-Shot Object Detection for Remote Sensing Images via Features Aggregation and Scale Attention

FSOD4RSI: Few-Shot Object Detection for Remote Sensing Images via Features Aggregation and Scale Attention

Few‐shot object detection based on global context and implicit knowledge decoupled head

Feature Alignment FPN for Oriented Object Detection in Remote Sensing Images

Contact Info

Product

Resources

About