LSTD: A Low-Shot Transfer Detector for Object Detection
This addresses the challenge of data scarcity in object detection for applications where fully-annotated datasets are limited, though it appears incremental as it builds on existing methods like SSD and Faster RCNN.
The paper tackles the problem of object detection with limited training data by proposing a low-shot transfer detector (LSTD) that leverages source-domain knowledge to build an effective target-domain detector with few examples, achieving state-of-the-art performance in challenging low-shot detection experiments.
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.