摘要
为了使模型能够使用少量样本就能对新类别进行学习,提出一种基于元学习的跨图像双重相似性小样本目标检测方法。提出支持特征自增强模块,丰富支持特征的表示能力。为了给查询特征分配支持信息,基于交叉注意力机制设计特征之间的跨图像相似性增强模块。在分类头中引入余弦相似度来计算查询向量和支持向量之间的距离度量用于修正网络的分类预测得分。在Pascal VOC数据集上的测试结果表明,改进后的算法能够有效地提高对新类别的目标检测能力,并将IoU阈值为0.5的平均均值精度mAP 0.5提升4.1百分点。
In order to enable the model to learn new categories using a small number of samples,the cross-image dual-similarity few-shot object detection method based on meta-learning is proposed.We proposed the feature self-enhancement module to enrich the representation capabilities of supporting features.In order to assign support information to query features,the cross image similarity enhancement module between features was designed based on cross-attention mechanism.Cosine similarity was introduced in the classification header to calculate the distance metric between the query vector and the support vector for correcting the classification prediction score of the network.The test results on the Pascal VOC dataset show that the improved algorithm can effectively improve the target detection ability for new categories,and improve the average mean accuracy mAP 0.5 with an IoU threshold of 0.5 increased by 4.1 percentage points.
作者
王浩然
朱煜
朱梓铭
林家骏
Wang Haoran;Zhu Yu;Zhu Ziming;Lin Jiajun(School of Information Science and Engineering,East China University of Science and Technology,Shanghai 200237,China)
出处
《计算机应用与软件》
北大核心
2026年第2期324-330,共7页
Computer Applications and Software
基金
上海市科学技术委员会科研计划项目(17DZ1100808)。
关键词
目标检测
小样本学习
深度学习
交叉注意力机制
Object detection
Few-shot learning
Deep learning
Cross-attention mechanism
