Giter Site home page Giter Site logo

N个连续的卷积层 about tood HOT 8 OPEN

fcjian avatar fcjian commented on June 1, 2024
N个连续的卷积层

from tood.

Comments (8)

fcjian avatar fcjian commented on June 1, 2024

@Lkevin20 不同层的特征具有不同的感受野大小,从而提取不同尺度的信息。

from tood.

Lkevin20 avatar Lkevin20 commented on June 1, 2024

作者你好,请问这个空间概率图和空间特征图是如何学习得到的?如果没有TAL策略是不是这个TAP就没有效果?那么为什么说T型头是一个独立的模块,可以在没有TAL的情况下很好地工作?

from tood.

Lkevin20 avatar Lkevin20 commented on June 1, 2024

这个交互特征又是如何得到的呢?辛苦作者解答一下谢谢

from tood.

fcjian avatar fcjian commented on June 1, 2024

作者你好,请问这个空间概率图和空间特征图是如何学习得到的?如果没有TAL策略是不是这个TAP就没有效果?那么为什么说T型头是一个独立的模块,可以在没有TAL的情况下很好地工作?

  1. 空间概率图的获取可以参考#13
  2. 没有TAL策略,T-head/TAL也有一定的效果,具体可以参考论文Table 1的实验。而有了TAL策略,可以使得T-head/TAL更好地工作,取得更高的性能。

from tood.

fcjian avatar fcjian commented on June 1, 2024

这个交互特征又是如何得到的呢?辛苦作者解答一下谢谢

通过N层共享的卷积:
image

from tood.

Lkevin20 avatar Lkevin20 commented on June 1, 2024

但是论文中说道空间概率图和偏移图的学习是在TAL中实现的所以有点不理解,论文中原话为“where conv1 and conv3 are two 1×1 conv layers for dimen-
sion reduction. The learning of M and O is performed by using the proposed Task Alignment Learning (TAL) which will be described next.”

from tood.

fcjian avatar fcjian commented on June 1, 2024

但是论文中说道空间概率图和偏移图的学习是在TAL中实现的所以有点不理解,论文中原话为“where conv1 and conv3 are two 1×1 conv layers for dimen- sion reduction. The learning of M and O is performed by using the proposed Task Alignment Learning (TAL) which will be described next.”

没有TAL策略,模型也可以学习到对检测性能有一定提高的M和O,比如传统的检测器选择物体中心区域的anchor作为正样本,此时学习到的M和O能够提高物体中心区域anchor预测的分类和定位的精度。
而如果采用TAL策略,此时学习到的M和O能够提高aligned anchor预测的分类和定位的精度,进一步提高分类和定位的对齐。
因此,学习到怎样的M/O跟所采用的学习策略有关。在TOOD里,正是采用了TAL来学习更能提高分类和定位对齐的M和O。

from tood.

Lkevin20 avatar Lkevin20 commented on June 1, 2024

这个TAL的分配策略和损失函数是在M和O上执行的吗?在原本的分类(P)和回归(B)上有体现吗?

from tood.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.