Abstract:Object detection is a critical field in computer vision focusing on accurately identifying and locating specific objects in images or videos. Traditional methods for object detection rely on large labeled training datasets for each object category, which can be time-consuming and expensive to collect and annotate. To address this issue, researchers have introduced few-shot object detection (FSOD) approaches that merge few-shot learning and object detection principles. These approaches allow models to quickly adapt to new object categories with only a few annotated samples. While traditional FSOD methods have been studied before, this survey paper comprehensively reviews FSOD research with a specific focus on covering different FSOD settings such as standard FSOD, generalized FSOD, incremental FSOD, open-set FSOD, and domain adaptive FSOD. These approaches play a vital role in reducing the reliance on extensive labeled datasets, particularly as the need for efficient machine learning models continues to rise. This survey paper aims to provide a comprehensive understanding of the above-mentioned few-shot settings and explore the methodologies for each FSOD task. It thoroughly compares state-of-the-art methods across different FSOD settings, analyzing them in detail based on their evaluation protocols. Additionally, it offers insights into their applications, challenges, and potential future directions in the evolving field of object detection with limited data.

What problem does this paper attempt to address?

This paper attempts to address the problem of reducing dependence on large-scale annotated datasets in object detection, especially in situations where it is impractical to collect a large amount of data in real-world applications. Specifically, the paper focuses on Few-Shot Object Detection (FSOD) and its different variants, including standard FSOD, Generalized FSOD (G-FSOD), Incremental FSOD (I-FSOD), Open-Set FSOD (O-FSOD), and Domain Adaptation FSOD (FSDAOD). These methods aim to enable the model to quickly adapt to new object categories using only a few annotated samples, thereby reducing the need for a large amount of annotated data. ### Specific Problems: 1. **Standard FSOD**: How to achieve object detection for new categories with only a few annotated samples? 2. **Generalized FSOD**: How to adapt to new categories while maintaining performance on known categories? 3. **Incremental FSOD**: How to gradually adapt to new categories without using data from old categories? 4. **Open-Set FSOD**: How to detect new object categories that were not seen in the training set? 5. **Domain Adaptation FSOD**: How to adapt the detector from one domain (e.g., source domain) to another domain (e.g., target domain) with only a few annotated data in the target domain? ### Background and Motivation: Traditional object detection methods rely on large-scale annotated datasets, which are often difficult to achieve in real-world applications because collecting and annotating a large amount of data is time-consuming and expensive. Additionally, training complex models with limited data can easily lead to overfitting problems. Therefore, Few-Shot Learning (FSL) has emerged, inspired by the human ability to quickly learn new concepts with only a few samples. FSOD applies this ability to object detection tasks to address the aforementioned issues. ### Application Scenarios: - **Medical Imaging**: Identifying rare diseases, enabling quick diagnosis and treatment. - **Wildlife Conservation**: Monitoring endangered species, supporting conservation efforts. - **Industrial Inspection**: Detecting defects or anomalies in manufacturing processes, improving quality control. - **Security and Surveillance**: Detecting suspicious activities or objects, enhancing security and response speed. - **Remote Sensing and Multispectral Imaging**: Addressing cross-domain generalization issues, expanding application scope. ### Contributions of the Paper: - **Comprehensive Review**: A detailed review of the research progress in FSOD and its different variants, analyzing the advantages and disadvantages of various methods. - **Classification and Comparison**: Classification and comparison of existing methods based on different training schemes and architectural layouts. - **Challenges and Future Directions**: Discussion of the challenges faced by FSOD tasks and proposing future research directions. In summary, this paper aims to provide researchers and practitioners with a comprehensive FSOD research framework to help them better understand and address the problem of data scarcity in object detection.

Beyond Few-shot Object Detection: A Detailed Survey

Few-Shot Object Detection: Research Advances and Challenges

Few-Shot Object Detection: A Comprehensive Survey

Recent Few-Shot Object Detection Algorithms: A Survey with Performance Comparison

Few-Shot Object Detection in Remote Sensing Image Interpretation: Opportunities and Challenges

A Survey of Deep Learning for Low-Shot Object Detection

A Survey of Self-Supervised and Few-Shot Object Detection

A Comparative Review of Recent Few-Shot Object Detection Algorithms

Few-Shot Object Detection in Unseen Domains

Few-shot Object Detection in Remote Sensing: Lifting the Curse of Incompletely Annotated Novel Objects

FSNA: Few-Shot Object Detection via Neighborhood Information Adaption and All Attention

Semantic Enhanced Few-shot Object Detection

Few-Shot Object Detection with Self-Adaptive Global Similarity and Two-Way Foreground Stimulator in Remote Sensing Images

Few-shot Weakly-Supervised Object Detection via Directional Statistics

Generalized few-shot object detection in remote sensing images

FSODv2: A Deep Calibrated Few-Shot Object Detection Network

Revisiting Few-Shot Object Detection with Vision-Language Models

FSODS: A Lightweight Metalearning Method for Few-Shot Object Detection on SAR Images

Few-Shot Object Detection in Remote-Sensing Images via Label-Consistent Classifier and Gradual Regression