Beyond Skip Connections -- TDM 特征融合模块设计思想

本文探讨了自顶向下调制在神经网络中的作用,通过VGG16和ResNet101网络的实验,展示了TDM模块如何增强小目标检测能力。在消融实验中发现,自顶向下路径对于传递上下文特征和选择相关细节至关重要,而侧向模块也对性能提升有显著影响。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

Top-down 的神经生物学基础

In the human visual pathway, once receptive field properties are tuned using feedforward processing,
top-down modulations are evoked by feedback and horizontal connections.

These connections modulate representations at multiple levels and are responsible for their selective combination.

We argue that the use of skip connections is a special case of this process, where the modulation is relegated to the final classifier, which directly tries to influence lower layer features and/or learn how to combine them.

[1] A. Gazzaley and A. C. Nobre. Top-down modulation: bridging selective attention and working memory. Trends in cognitive sciences, 2012.
[2] C. D. Gilbert and M. Sigman. Brain st

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包

打赏作者

Gallant Hu

你的鼓励将是我创作的最大动力

¥1 ¥2 ¥4 ¥6 ¥10 ¥20
扫码支付:¥1
获取中
扫码支付

您的余额不足,请更换扫码支付或充值

打赏作者

实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值