3 holeung

尚未进行身份认证

东南大学

等级
TA的排名 1w+

CVPR&ICCV部分注意力机制最新工作调研

2019-09-28 09:20:14

[论文笔记] (CVPR2019) YOLACT: Real-time Instance Segmentation

2019-08-15 00:25:16

[论文笔记] (CVPR2019) FastFCN:Rethinking Dilated Convolution in the Backbone for Semantic Segmentation

2019-08-15 00:22:13

[论文笔记] (CVPR2019) DeeperLab: Single-Shot Image Parser

2019-08-15 00:15:07

[论文笔记] (CVPR2019) DFANet: Deep Feature Aggregation for Real-Time Semantic Segmentation

2019-08-15 00:08:19

雨夜

2019.8.3 于苏州

2019-08-04 09:21:25

[论文笔记] (CVPR2019) Structured Knowledge Distillation for Semantic Segmentation

[论文笔记] (CVPR2019) Structured Knowledge Distillation for Semantic Segmentation

2019-08-04 09:09:06

[论文笔记] (CVPR2019) Panoptic Feature Pyramid Networks

[论文笔记] (CVPR2019) Panoptic Feature Pyramid Networks

2019-08-04 09:06:03

[论文笔记] (CVPR2019) Mask Scoring R-CNN

[论文笔记] (CVPR2019) Mask Scoring R-CNN

2019-08-04 09:03:04

[论文笔记] (CVPR2019) Attention-guided Unified Network for Panoptic Segmentation

[论文笔记] (CVPR2019) Attention-guided Unified Network for Panoptic Segmentation

2019-08-04 09:00:12

[论文笔记] (CVPR2019) UPSNet: A Unified Panoptic Segmentation Network

[论文笔记] (CVPR2019) UPSNet: A Unified Panoptic Segmentation Network

2019-08-04 08:53:53

[论文笔记] NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection(CVPR19)

[论文笔记] NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection(CVPR19)

2019-08-04 08:51:11

[论文笔记] Weight Agnostic Neural Networks

Experiment

2019-08-04 08:50:21

[论文笔记] Panoptic Segmentation with a Joint Semantic and Instance Segmentation Network

[论文简记] Panoptic Segmentation with a Joint Semantic and Instance Segmentation Network

2019-08-04 08:49:11

[论文笔记] (CVPR2019) An End-to-End Network for Panoptic Segmentation

[论文笔记] (CVPR2019) An End-to-End Network for Panoptic Segmentation```

2019-08-04 08:46:18

[论文笔记] (CVPR2019) Panoptic Segmentation

[论文笔记] (CVPR2019) Panoptic Segmentation

2019-08-04 08:41:58

[论文简记] (CVPR19) Decoders Matter for Semantic Segmentation

该文章提出了一种不同于双线性插值的上采样方法,能够更好的建立每个像素之间预测的相关性。得益于这个强大的上采样方法,模型能够减少对特征图分辨率的依赖,能极大的减少运算量。该工作在 PASCAL VOC 数据集上达到了 88.1% 的 mIOU,超过了 DeeplabV3 + 的同时只有其 30% 的计算量。在之前的语义分割方法中,双线性插值通常作为其最后一步来还原特征图的分辨率,由于非线性...

2019-07-29 10:46:25

[论文极简笔记] Searching for A Robust Neural Architecture in Four GPU Hours (CVPR19)

GDAS将整个搜索空间用一个有向无环图(DAG)来表示。针对这个DAG,GDAS设计了一个可微的采样器,GDAS在训练集上优化DAG内每个网络结构的参数,在验证集上优化这个可微的神经网络采样器。实验表明,在一个GPU上,通过几个小时的搜索时间,GDAS就可以在CIFAR-10数据集上找到一个高性能的网络结构。...

2019-07-05 21:18:29

[论文极简笔记] Auto-DeepLab: Hierarchical Neural Architecture Search for Semantic Image Segmentation

2019-07-05 21:16:35

[论文极简笔记] Partial Order Pruning: for Best Speed/Accuracy Trade-off in Neural Architecture Search

CVPR2019Contribute:This assumption may not hold for very deep networks that contain hundreds of layers , but it is generally true for the efficient architectures of our concern,Partial O...

2019-07-05 21:10:09

查看更多

勋章 我的勋章
  • 专栏达人
    专栏达人
    授予成功创建个人博客专栏的用户。专栏中添加五篇以上博文即可点亮!撰写博客专栏浓缩技术精华,专栏达人就是你!
  • 持之以恒
    持之以恒
    授予每个自然月内发布4篇或4篇以上原创或翻译IT博文的用户。不积跬步无以至千里,不积小流无以成江海,程序人生的精彩需要坚持不懈地积累!
  • 勤写标兵Lv2
    勤写标兵Lv2
    授予每个自然周发布4篇到6篇原创IT博文的用户。本勋章将于次周周三上午根据用户上周的博文发布情况由系统自动颁发。