LSTD: A Low-Shot Transfer Detector for Object Detection

By Hao Chen, Yali Wang, Guoyou Wang, Yu Qiao. 2018

392 citations recorded by the source corpus.

Recent advances in object detection are mainly driven by deep learning with large-scale detection benchmarks. However, the fully-annotated training set is often limited for a target detection task, which may deteriorate the performance of deep detectors. To address this challenge, we propose a novel low-shot transfer detector (LSTD) in this paper, where we leverage rich source-domain knowledge to construct an effective target-domain detector with very few training examples. The main contributions are described as follows. First, we design a flexible deep architecture of LSTD to alleviate transfer difficulties in low-shot detection. This architecture can integrate the advantages of both SSD and Faster RCNN in a unified deep framework. Second, we introduce a novel regularized transfer learning framework for low-shot detection, where the transfer knowledge (TK) and background depression (BD) regularizations are proposed to leverage object knowledge respectively from source and target domains, in order to further enhance fine-tuning with a few target images. Finally, we examine our LSTD on a number of challenging low-shot detection experiments, where LSTD outperforms other state-of-the-art approaches. The results demonstrate that LSTD is a preferable deep detector for low-shot scenarios.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Radiology, Nuclear Medicine and Imaging, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Radiology, Nuclear Medicine and Imaging

In Meta & Continual Learning

Citations in the corpus

Learning Open-World Object Proposals without Learning to Classify (2021, background): “This line of research is also related to few-shot [10, 31, 67, 68, 71] and zero-shot detection [6, 20, 43, 70] methods, which attempt to detect novel classes given only a few samples or class descriptions.”
Few-Shot Object Detection via Feature Reweighting (2018, background): “[4] studies a similar setting but only in a transfer learning context, where the target domain images only contains novel classes without base classes.”
RepMet: Representative-Based Metric Learning for Classification and Few-Shot Object Detection (2018, methodology): “To the best of our knowledge, the only few-shot detection benchmark available to-date is reported in the LSTD work [5] by Chen et al., who proposed to approach fewshot detection by a regularized fine-tuning.”
Few-Shot Adaptive Faster R-CNN (2019, background): “[8] proposed to learn an object detector with a large pool of unlabeled images and only a few annotated images per category, termed few-shot object detection (FSOD); [4] tackled the setting of few-shot object detection with a low-shot transfer detector (LSTD) coupled with designed regularization.”
Few-Shot Object Detection and Viewpoint Estimation for Objects in the Wild (2020, neutral): “[15] proposes to transfer a pre-trained detector to the few-shot task, while Karlinsky et al.”
Recent Advances in Deep Learning for Object Detection (2019, background): “(LSTD) [254] addresses low-shot detection based on transfer learning which transfers the knowledge form large annotated external datasets to the target set by knowledge regularization.”
One-Shot Instance Segmentation (2018, methodology): “The low-shot transfer detector (LSTD) [7] fine-tunes an object detector on a transfer task with new categories using two novel regularization terms: one for background depression and one for knowledge transfer from the source domain to the target domain.”
Meta-Learning to Detect Rare Objects (2019, methodology): “scale, few-shot detection benchmark in [6].”
Any-Shot Object Detection (2020, neutral): “3 LSTD [6] 74.”
Deep Learning for Generic Object Detection: A Survey (2018, background): “n learn a visual concept quickly from very few given examples and can often generalize well [16, 144,71]. Therefore, the ability to learn from only few examples, few shot detection, is very appealing [30,61,75,129,144,228, 237]. Even more constrained, zero shot object detection localizes and recogni”
Few-Shot Object Detection With Attention-RPN and Multi-Relation Detector (2019, methodology): “In Table 4, we compare our results with those of LSTD [9] and RepMet [61] on the challenging ImageNet based 50-way 5-shot detection scenario.”
Overcoming Classifier Imbalance for Long-Tail Object Detection With Balanced Group Softmax (2020, background): “Compared with balanced distribution targeted object detection [11, 31, 1], and few-shot object detection [19, 3, 38, 10], the challenging and practical long-tail object detection problem is still underexplored.”
The Devil is in Classification: A Simple Framework for Long-tail Object Detection and Instance Segmentation (2020, background): “For example, AP1 corresponds to the commonly referred few-shot object detection regime [8], [23], [25].”
The Devil is in Classification: A Simple Framework for Long-tail Instance Segmentation (2020, background): “For example, AP 1 corresponds to the commonly referred few-shot object detection regime [8], [23], [25].”
DeFRCN: Decoupled Faster R-CNN for Few-Shot Object Detection (2021, methodology): “As one of the research branches, fewshot object detection (FSOD) is a much more challenging task than both few-shot classification and object detection [5, 19, 46, 51, 53].”
Multi-Scale Positive Sample Refinement for Few-Shot Object Detection (2020, methodology): “Different from the original implementation in [31], we employ a class-agnostic regression task in the detection head, which is the same as [4].”
FGN: Fully Guided Network for Few-Shot Instance Segmentation (2020, background): “Some works address this problem from the perspectives of selfpaced learning [5] or transfer learning [2].”