Collect some papers about few-shot object detection for computer vision. Additionally, we briefly introduce the commonly used datasets for few-shot object detection.
| Title | Venue | |
|---|---|---|
| A Survey of Deep Learning for Low-Shot Object Detection | ArXiv 2022 | |
| A Unified Framework for Attention-Based Few-Shot Object Detection | ArXiv 2022 |
| Title | Venue | Dataset | CODE | |
|---|---|---|---|---|
| Kernelized Few-shot Object Detection with Efficient Integral Aggregation | CVPR 2022 | PASCAL VOC & MS COCO | CODE | |
| Label, Verify, Correct: A Simple Few Shot Object Detection Method | CVPR 2022 | PASCAL VOC & MS COCO | CODE | |
| Few-Shot Object Detection with Fully Cross-Transformer | CVPR 2022 | PASCAL VOC & MS COCO | - | |
| Meta Faster R-CNN: Towards Accurate Few-Shot Object Detection with Attentive Feature Alignment | AAAI 2022 | PASCAL VOC & MS COCO | CODE | |
| Few-Shot Object Detection by Attending to Per-Sample-Prototype | WACV 2022 | PASCAL VOC & MS COCO | - |
| Title | Venue | Dataset | CODE | |
|---|---|---|---|---|
| Generalized and Discriminative Few-Shot Object Detection via SVD-Dictionary Enhancement | NeurIPS 2021 | PASCAL VOC & MS COCO | CODE | |
| Few-Shot Object Detection via Association and DIscrimination | NeurIPS 2021 | PASCAL VOC & MS COCO | CODE | |
| Adaptive Image Transformer for One-Shot Object Detection | CVPR 2021 | PASCAL VOC & MS COCO | CODE | |
| Dense Relation Distillation with Context-aware Aggregation for Few-Shot Object Detection | CVPR 2021 | PASCAL VOC & MS COCO | CODE | |
| Generalized Few-Shot Object Detection without Forgetting | CVPR 2021 | PASCAL VOC & MS COCO | CODE | |
| Beyond Max-Margin: Class Margin Equilibrium for Few-shot Object Detection | CVPR 2021 | PASCAL VOC & MS COCO | CODE | |
| Few-Shot Object Detection via Classification Refinement and Distractor Retreatment | CVPR 2021 | PASCAL VOC & MS COCO | - | |
| Transformation Invariant Few-Shot Object Detection | CVPR 2021 | PASCAL VOC & MS COCO | - | |
| UniT: Unified Knowledge Transfer for Any-shot Object Detection and Segmentation | CVPR 2021 | PASCAL VOC & MS COCO | CODE | |
| FAPIS: A Few-shot Anchor-free Part-based Instance Segmenter | CVPR 2021 | MS COCO | CODE | |
| Semantic Relation Reasoning for Shot-Stable Few-Shot Object Detection | CVPR 2021 | PASCAL VOC & MS COCO | - | |
| FSCE: Few-Shot Object Detection via Contrastive Proposal Encoding | CVPR 2021 | PASCAL VOC & MS COCO | CODE | |
| Accurate Few-shot Object Detection with Support-Query Mutual Guidance and Hybrid Loss | CVPR 2021 | PASCAL VOC & MS COCO | - | |
| Hallucination Improves Few-Shot Object Detection | CVPR 2021 | PASCAL VOC & MS COCO | CODE | |
| Query Adaptive Few-Shot Object Detection with Heterogeneous Graph Convolutional Networks | ICCV 2021 | PASCAL VOC & MS COCO | CODE | |
| Universal-Prototype Enhancing for Few-Shot Object Detection | ICCV 2021 | PASCAL VOC & MS COCO | CODE | |
| DeFRCN: Decoupled Faster R-CNN for Few-Shot Object Detection | ICCV 2021 | PASCAL VOC & MS COCO | CODE | |
| Meta-DETR: Few-Shot Object Detection via Unified Image-Level Meta-Learning | ArXiv 2021 | PASCAL VOC & MS COCO | CODE |
| Title | Venue | Dataset | CODE | |
|---|---|---|---|---|
| Restoring Negative Information in Few-Shot Object Detection | NeurIPS 2020 | PASCAL VOC & ImageNet-LOC | CODE | |
| Context-Transformer: Tackling Object Confusion for Few-Shot Detection | AAAI 2020 | PASCAL VOC & MS COCO | CODE | |
| Few-Shot Object Detection with Attention-RPN and Multi-Relation Detector | CVPR 2020 | MS COCO & ImageNet-LOC | CODE | |
| Few-Shot Object Detection and Viewpoint Estimation for Objects in the Wild | ECCV 2020 | PASCAL VOC & MS COCO | CODE | |
| Multi-Scale Positive Sample Refinement for Few-Shot Object Detection | ECCV 2020 | PASCAL VOC & MS COCO | CODE | |
| META-RCNN: META LEARNING FOR FEW-SHOT OBJECT DETECTION | ICLR 2020 | PASCAL VOC & ImageNet-LOC | - | |
| Frustratingly Simple Few-Shot Object Detection | ICML 2020 | PASCAL VOC & MS COCO & LVIS | CODE |
| Title | Venue | Dataset | CODE | |
|---|---|---|---|---|
| Few-shot Object Detection via Feature Reweighting | ICCV 2019 | PASCAL VOC & MS COCO | CODE | |
| RepMet: Representative-based metric learning for classification and few-shot object detection | CVPR 2019 | PASCAL VOC & ImageNet-LOC | CODE | |
| Meta R-CNN : Towards General Solver for Instance-level Low-shot Learning | ICCV 2019 | PASCAL VOC & MS COCO | CODE | |
| One-Shot Object Detection with Co-Attention and Co-Excitation | NeurIPS 2019 | PASCAL VOC & MS COCO | CODE |
| Title | Venue | Dataset | CODE | |
|---|---|---|---|---|
| Few-example object detection with model communication | TPAMI 2018 | PASCAL VOC & MS COCO & ILSVRC | - | |
| LSTD: A Low-Shot Transfer Detector for Object Detection | AAAI 2018 | PASCAL VOC & MS COCO & ImageNet2015 | - |
Most few-shot object detection papers usually follow the experimental setup in 《Meta R-CNN : Towards General Solver for Instance-level Low-shot Learning》 and 《Few-shot Object Detection via Feature Reweighting》 and conduct experiments on PASCAL VOC and MS COCO datasets.
For Pascal VOC, the model is trained on the trainval sets of VOC 2007 and 2012, and tested on VOC 2007 test set. The dataset is partitioned into three base/novel splits, where 5 categories are selected as novel classes and others are base classes. The novel classes in each split are as follows:
| Split | Novel Class |
|---|---|
| 1 | bird, bus, cow, mbike, sofa |
| 2 | aero, bottle, cow, horse, sofa |
| 3 | boat, cat, mbike, sheep, sofa |
MS COCO with 80 object categories including the 20 categories in PASCAL VOC. The 20 classes included in PASCAL VOC are set as novel classes, then the rest 60 classes in COCO are as base classes. The union of 80k train images and a 35k subset of validation images (trainval35k) are used for training, and the evaluation is based on the remaining 5k val images (minival).
Thanks to Yongqiang Mao for the ideas and templates.