摘要
为了完善目标检测中的小目标、遮挡等造成的漏检问题,提出融合上下文的强化学习算法。在特征提取阶段,采用U型网络框架,并通过引入上下文机制,推测判定可能存在的特征,利用向量拼接的方式得到最终特征。在目标检测阶段,引入强化学习Soft Actor-Critic模型来提升检测精度。实验证明,该方法对目标检测精确度有了明显的提升,有效地提高了对小目标的检测能力。
In order to improve the problem of missing detection caused by small target and occlusion in target detection,a context-fusion reinforcement learning algorithm is proposed.In the feature extraction stage,the U-shaped network framework was used,and the context mechanism was introduced to infer and determine the possible features.The final features were obtained by vector splicing.In the target detection stage,the soft Actor-critic model was introduced to improve the detection accuracy.Experimental results show that this method can improve the accuracy of target detection and effectively improve the detection ability of small targets.
作者
曹立春
智敏
Cao Lichun;Zhi Min(College of Computer Science and Technology,Inner Mongolia Normal University,Hohhot 010022,Inner Mongolia,China)
出处
《计算机应用与软件》
北大核心
2023年第5期221-226,共6页
Computer Applications and Software
基金
内蒙古自然科学基金项目(2018MS06008)。
