蔡文迪的博客

FCOS Fully Convolutional One-Stage Object Detection

论文阅读笔记

阅读笔记论文相关信息题目：FCOS: Fully Convolutional One-Stage Object Detection 作者：Zhi Tian, Chunhua Shen, Hao Chen, Tong He 出处：https://arxiv.org/pdf/1904.01355.pdf 摘要本文提出了一种全卷积的单阶段目标检测（FCOS）方法，以逐像素预测...

Posted by Kiri on May 10, 2019

YOLOv3 An Incremental Improvement

论文阅读笔记

阅读笔记论文相关信息题目：YOLOv3: An Incremental Improvement 作者：Joseph Redmon, Ali Farhadi 来源：https://arxiv.org/pdf/1804.02767.pdf 摘要本文提出了一些YoLo的改进方法。文中做了一些设计上的小变化，使之表现得更好。在$320\times 320$输入下YoLov3的检...

Posted by Kiri on May 5, 2019

Multi-task Convolutional Neural Network for Patient Detection and Skin Segmentation in Continuous Non-contact Vital Sign Monitoring

论文阅读笔记

#阅读笔记论文相关信息题目：Multi-task Convolutional Neural Network for Patient Detection and Skin Segmentation in Continuous Non-contact Vital Sign Monitoring 作者：Sitthichok Chaichulee, Mauricio Villarroel,...

Posted by Kiri on April 12, 2019

Deep Learning Based Hand Detection in Cluttered Environment Using Skin Segmentation

论文阅读笔记

阅读笔记论文相关信息题目：Deep Learning Based Hand Detection in Cluttered Environment Using Skin Segmentation 作者：Kankana Roy, Aparna Mohanty, Rajiv R. Sahay 出处：Proceedings of the IEEE International Confer...

Posted by Kiri on April 12, 2019

Res2Net A New Multi-scale Backbone Architecture

论文阅读笔记

阅读笔记论文相关信息题目：Res2Net: A New Multi-scale Backbone Architecture 来源：https://arxiv.org/abs/1904.01169 作者：Shang-Hua Gao, Ming-Ming Cheng, Kai Zhao, Xin-Yu Zhang, Ming-Hsuan Yang, Philip Torr 摘要...

Posted by Kiri on April 12, 2019

DSSD Deconvolutional Single Shot Detector

论文阅读笔记

阅读笔记论文相关信息：题目：DSSD: Deconvolutional Single Shot Detector 来源：https://arxiv.org/abs/1701.06659（未在期刊或者会议发表）作者：Cheng Yang Fu, Wei Liu, Ananth Ranga, Ambrish Tyagi, Alexander C. Berg 摘要：本文首先...

Posted by Kiri on March 23, 2019

Feature Pyramid Networks for Object Detection

论文阅读笔记

阅读笔记题目：Feature Pyramid Networks for Object Detection 来源：CVPR2017 背景：Feature Pyramid是检测多尺度物体的识别系统中一个基本的组成部分。但近几年来基于深度学习的目标检测方法避免了Pyramid表示，很大程度上因为它受到计算和存储的制约。本文主要介绍如何使用网络内的特征金字塔代替特征化的图像金字塔，而不牺...

Posted by Kiri on March 15, 2019

SSD Single Shot MultiBox Detector

论文阅读笔记

阅读笔记题目：SSD: Single Shot MultiBox Detector SSD特点：使用单一深度神经网络网络提取了不同尺度的特征图，自然的处理各种尺度的物体 Model: 标准的卷积神经网络结构用于整个网络的骨架部分，然后添加一些辅助结构进行目标检测。采用多尺度特征图用于检测采用卷积进行检测对于形状...

Posted by Kiri on March 8, 2019

PyTorch——Load Dataset

PyTorch Load Dataset import torch import numpy as np import cv2 import torch.utils.data as data import os from skimage import io, transform import matplotlib.pyplot as plt import matplotlib.ima...

Posted by Kiri on February 25, 2019

Anchors

基于候选区域的目标检测算法

Anchors 在Feature Map上选用用一个$3\times 3$的滑窗，每个滑窗生成9个anchors（具有相同中心点$(x_a,y_a）$，但具有3种不同的长宽比（aspect ratios）和3种不同的尺寸（scales），计算的是相对于原始图片的尺寸）对每个anchor计算与ground-truth bbox的重叠比IOU if IoU > 0.7 p = ...

Posted by Kiri on December 28, 2018

Kiri Blog

FCOS Fully Convolutional One-Stage Object Detection

论文阅读笔记

YOLOv3 An Incremental Improvement

论文阅读笔记

Multi-task Convolutional Neural Network for Patient Detection and Skin Segmentation in Continuous Non-contact Vital Sign Monitoring

论文阅读笔记

Deep Learning Based Hand Detection in Cluttered Environment Using Skin Segmentation

论文阅读笔记

Res2Net A New Multi-scale Backbone Architecture

论文阅读笔记

DSSD Deconvolutional Single Shot Detector

论文阅读笔记

Feature Pyramid Networks for Object Detection

论文阅读笔记

SSD Single Shot MultiBox Detector

论文阅读笔记

PyTorch——Load Dataset

Anchors

基于候选区域的目标检测算法

FEATURED TAGS

ABOUT ME

FRIENDS