论文标题
r（det）^2：对象检测的随机决策路由
R(Det)^2: Randomized Decision Routing for Object Detection
论文作者
论文摘要
在对象检测范式中，决策头是重要的部分，它会严重影响检测性能。然而，如何设计高性能决策负责人仍然是一个悬而未决的问题。在本文中，我们提出了一种新颖的方法，以端到端的学习方式将决策树和深层神经网络相结合，以进行对象检测。首先，我们通过将软决策树插入神经网络来解开决策选择和预测值。为了促进有效的学习，我们建议使用节点选择性和关联损失的随机决策路由，这可以同时提高功能代表性学习和网络决策。其次，我们开发了用狭窄分支的对象检测的决策头，以生成路由概率和掩模，以便从不同节点获得不同的决策。我们将这种方法称为用于对象检测的随机决策路由，缩写为r（det）$^2 $。 MS-Coco数据集的实验表明，R（det）$^2 $可有效提高检测性能。配备现有检测器，可实现$ 1.4 \ sim 3.6 $ \％AP改进。
In the paradigm of object detection, the decision head is an important part, which affects detection performance significantly. Yet how to design a high-performance decision head remains to be an open issue. In this paper, we propose a novel approach to combine decision trees and deep neural networks in an end-to-end learning manner for object detection. First, we disentangle the decision choices and prediction values by plugging soft decision trees into neural networks. To facilitate effective learning, we propose randomized decision routing with node selective and associative losses, which can boost the feature representative learning and network decision simultaneously. Second, we develop the decision head for object detection with narrow branches to generate the routing probabilities and masks, for the purpose of obtaining divergent decisions from different nodes. We name this approach as the randomized decision routing for object detection, abbreviated as R(Det)$^2$. Experiments on MS-COCO dataset demonstrate that R(Det)$^2$ is effective to improve the detection performance. Equipped with existing detectors, it achieves $1.4\sim 3.6$\% AP improvement.
