论文笔记 - 随笔分类 - StoneclutterX

论文笔记-StableCascade

摘要：资料 paper： a. arxiv：http://arxiv.org/abs/2306.00637 b. ICLR版本：https://openreview.net/pdf?id=gU58d5QeGv stability.ai版本（模型效果更好？） a. code：https://github.c 阅读全文

posted @ 2024-02-28 14:57 StoneclutterX 阅读(172) 评论(0) 推荐(0)

论文笔记-PolarMask: Single Shot Instance Segmentation with Polar Representation

摘要：paper: PolarMask: Single Shot Instance Segmentation with Polar Representation code: https://github.com/xieenze/PolarMask 作者知乎解答 PolarMask是CVPR2020的ora 阅读全文

posted @ 2020-05-28 12:07 StoneclutterX 阅读(1149) 评论(0) 推荐(0)

论文笔记-IntRA-KD: Inter-Region Affinity Distillation for Road Marking Segmentation

摘要：paper: "Inter Region Affinity Distillation for Road Marking Segmentation" code: "Codes for IntRA KD" Abstract 1. IntRA KD是商汤&港中文&南洋理工在CVPR2020的一篇road 阅读全文

posted @ 2020-05-23 17:18 StoneclutterX 阅读(1985) 评论(0) 推荐(0)

论文笔记-ERFNet: Efficient Residual Factorized ConvNet for Real-time Semantic Segmentation

摘要：paper: "ERFNet: Efficient Residual Factorized ConvNet for Real time Semantic Segmentation" code: "PyTorch" Abstract 1. ERFNet可以看作是对ResNet结构的又一改变，同时也是对阅读全文

posted @ 2020-05-23 11:17 StoneclutterX 阅读(3034) 评论(0) 推荐(0)

论文笔记-ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation

摘要：paper: "ENet: A Deep Neural Network Architecture for Real Time" code: "caffe" Abstract 1. ENet是16年初的一篇工作了，能够达到实时的语义分割，包括在嵌入式设备NVIDIA TX1，同时还能够保证网络的效果。阅读全文

posted @ 2020-05-22 16:49 StoneclutterX 阅读(628) 评论(0) 推荐(0)

论文笔记-FCOS: Fully Convolutional One-Stage Object Detection

摘要：paper: "FCOS: Fully Convolutional One Stage Object Detection" code: "https://github.com/tianzhi0549/FCOS/" Abstract 1. FCOS是ICCV2019的一篇anchor free的one 阅读全文

posted @ 2020-05-20 22:13 StoneclutterX 阅读(915) 评论(0) 推荐(0)

论文笔记-ResNeSt-Split-Attention Networks

摘要：paper: "ResNeSt: Split Attention Networks" code: "ResNeSt Mxnet&PyTorch版本" Abstract 1. ResNeSt是对ResNext的又一个魔改的变种，亚马逊李沐团队的作品，其中S代表split，可以理解为按channel切分阅读全文

posted @ 2020-05-18 22:30 StoneclutterX 阅读(1577) 评论(0) 推荐(0)

论文笔记-HRNet-Deep High-Resolution Representation Learning for Visual Recognition

摘要：paper: "Deep High Resolution Representation Learning for Visual Recognition" code: "HRNet" Abstract 1. HRNet，这里用的是PAMI2020的工作，整合了human pose estimation 阅读全文

posted @ 2020-05-11 19:47 StoneclutterX 阅读(2955) 评论(0) 推荐(0)

论文笔记-OCR-Object-Contextual Representations for Semantic Segmentation

摘要：paper: "Object Contextual Representations for Semantic Segmentation" code: "PyTorch" Abstract 1. OCR是MSRA和中科院的一篇语义分割工作，结合每一类的类别语义信息给每个像素加权，再和原始的pixel特阅读全文

posted @ 2020-05-10 14:31 StoneclutterX 阅读(5103) 评论(4) 推荐(2)

论文笔记-经典网络之ResNet再读

摘要：paper: Deep Residual Learning for Image Recognition Identity Mappings in Deep Residual Networks code: caffe slides ResNet ResNet，一代经典，将深度网络结构设计以及各个视觉任阅读全文

posted @ 2020-05-09 22:16 StoneclutterX 阅读(568) 评论(1) 推荐(1)

论文笔记-PSPNet-Pyramid Scene Parsing Network

摘要：paper: "Pyramid Scene Parsing Network" code: "caffe" ， "PyTorch" "官方blog" Abstract 1. PSPNet是港中文和商汤在CVPR2017上的一篇工作，将pyramid pooling引入到场景解析的分割任务中，取得很好的阅读全文

posted @ 2020-05-08 21:13 StoneclutterX 阅读(1061) 评论(0) 推荐(0)

论文笔记-U-Net: Convolutional Networks for Biomedical Image Segmentation

摘要：paper: "U Net: Convolutional Networks for Biomedical Image Segmentation" code: "PyTorch" ，官方公布的matlab的caffe版本如果需要看网络结构，可以用这个prototxt文件 "unet prototxt 阅读全文

posted @ 2020-05-08 15:28 StoneclutterX 阅读(1262) 评论(0) 推荐(0)

论文笔记-Strip Pooling: Rethinking Spatial Pooling for Scene Parsing

摘要：paper: Strip Pooling: Rethinking Spatial Pooling for Scene Parsing code: https://github.com/Andrew-Qibin/SPNet Abstract CVPR2020的一篇文章，通过改进空间池化层来优化场景分割阅读全文

posted @ 2020-04-22 01:12 StoneclutterX 阅读(2398) 评论(1) 推荐(1)

论文笔记-Learning Lightweight Lane Detection CNNs by Self Attention Distillation

摘要：paper: "Learning Lightweight Lane Detection CNNs by Self Attention Distillation" code: "https://github.com/cardwing/Codes for Lane Detection" Abstract 阅读全文

posted @ 2020-04-15 19:31 StoneclutterX 阅读(1627) 评论(0) 推荐(1)

论文笔记-CenterNet: Objects as points without anchor

摘要：paper: Objects as Points code: https://github.com/xingyizhou/CenterNet Abstract CenterNet是anchor-free的目标检测方法，用预测出的点表示一个object，同时回归出目标的size和offset。其性能表阅读全文

posted @ 2020-04-15 18:00 StoneclutterX 阅读(426) 评论(0) 推荐(0)

论文笔记-Focal Loss for Dense Object Detection

摘要：paper: "Focal Loss for Dense Object Detection" Abstract 1. 目前主流的目标检测方法，one stage相较于two stage的一个痛点就是样本分布的不平衡（目标和背景），导致one stage虽然速度上有优势，但是精度却有所降低。因此，本文阅读全文

posted @ 2020-04-15 17:55 StoneclutterX 阅读(338) 评论(0) 推荐(0)

论文笔记-YOLO系列

摘要：Yolo V1 paper: "You Only Look Once: Unified, Real Time Object Detection" slide： "官方slide介绍，很详细" Abstract 1. 如文章名字所述，其完成一次目标检测，只需要look once，也就是不像两阶段的网络阅读全文

posted @ 2020-04-15 17:28 StoneclutterX 阅读(690) 评论(0) 推荐(0)

论文笔记-EfficientDet

摘要：paper: "EfficientDet: Scalable and Efficient Object Detection" code "官方tf" "民间PyTroch" Abstract 1. 谷歌大脑团队提出的新的目标检测算法，性能极佳，EfficientDet D0的mAP就超过了yolov 阅读全文

posted @ 2020-04-15 17:18 StoneclutterX 阅读(544) 评论(0) 推荐(0)

论文笔记-Understanding Convolution for Semantic Segmentation

摘要：图森和CMU的合作工作。论文链接[https://arxiv.org/abs/1702.08502](https://arxiv.org/abs/1702.08502) 主要提出DUC（dense upsampling convolution）和HDC（hybrid dilated convolu 阅读全文

posted @ 2018-10-12 21:45 StoneclutterX 阅读(1881) 评论(0) 推荐(0)

StoneclutterX

随笔分类 - 论文笔记

公告