- 博客(27)
- 资源 (91)
- 收藏
- 关注
原创 CNN网络加速--Performance Guaranteed Network Acceleration via High-Order Residual Quantization
Performance Guaranteed Network Acceleration via High-Order Residual Quantization ICCV2017本文是对 XNOR-Networks 的改进,将CNN网络层的输入 进行高精度二值量化,从而实现高精度的二值网络计算,XNOR-Networks 也是对每个CNN网络层的权值和输入进行二值化,这样整个CNN计算都是二值化
2017-08-31 15:18:24 1993 1
原创 目标检测--RON: Reverse Connection with Objectness Prior Networks for Object Detection
RON: Reverse Connection with Objectness Prior Networks for Object Detection CVPR2017 https://github.com/taokong/RON本文可以看作是对 SSD 的改进, SSD 对不同尺度特征图进行独立的检测,这里我们 reverse connection block 将相邻的特征图联系起来。同时使用
2017-08-30 15:55:11 5147 2
原创 去水印--《On the Effectiveness of Visible Watermarks》
On the Effectiveness of Visible Watermarks CVPR2017大牛 William T. Freeman 去了 Google Research 本文没有使用深度学习,使用传统的优化算法来解决水印去除。首先看图有个感性认识吧 基于单张图像去除水印的难度还是很大的。这里我们在网上搜集使用同一个水印的大量图像,基于这些图像,我们估计出 Watermark (W
2017-08-30 10:57:52 10704 6
原创 CNN阴影去除--DeshadowNet: A Multi-context Embedding Deep Network for Shadow Removal
DeshadowNet: A Multi-context Embedding Deep Network for Shadow Removal CVPR2017本文使用深度学习CNN网络来进行阴影去除,最大的特色就是全自动的端对端的实现阴影去除。 automatic and end-to-end deep neural network (DeshadowNet)阴影去除也算是一个老大难问题了,目前存
2017-08-29 13:49:15 8449 11
原创 CNN边缘检测--Richer Convolutional Features for Edge Detection
Richer Convolutional Features for Edge Detection CVPR2017 Caffe:https://github.com/yun-liu/rcf本文针对边缘检测问题,基于 VGG16 网络设计了一个 richer convolutional features (RCF) 用于边缘检测,效果目前是很好的。首先来看看 VGG16不同卷积层的特征输出 3
2017-08-28 16:54:39 13627 1
原创 目标检测分割--BlitzNet: A Real-Time Deep Network for Scene Understanding
BlitzNet: A Real-Time Deep Network for Scene Understanding ICCV2017 Project: http://thoth.inrialpes.fr/research/blitznet/ Code: https://github.com/dvornikita/blitznet本文在 SSD 基础上改进提出 BlitzNet,使其
2017-08-25 16:11:50 4668
原创 语义分割--Deep Dual Learning for Semantic Image Segmentation
Deep Dual Learning for Semantic Image Segmentation ICCV2017针对语义分割问题,本文提出了一个 dual image segmentation (DIS)系统 利用一部分 per-pixel labelmaps的训练样本和 一部分 image-level tags 的样本 进行联合训练,得到较好的分割结果。本文定义了一些符号: I 输入图
2017-08-25 11:09:02 3433
原创 分割候选区域--FastMask: Segment Multi-scale Object Candidates in One Shot
FastMask: Segment Multi-scale Object Candidates in One Shot CVPR2017 https://github.com/voidrank/FastMask本文针对检测和分割问题提出 FastMask 实现 segment multi-scale objects in one shot 这里的 one shot ( original im
2017-08-24 14:30:43 1851
原创 行人检测--What Can Help Pedestrian Detection?
What Can Help Pedestrian Detection? CVPR2017本文主要分析 extra features 对于 基于CNN 的行人检测有什么帮助,设计了一个行人检测网络 HyperLearner 可以有效利用这些 extra features这里的 extra features 主要指 various channel features 行人检测中存在什么问题了?
2017-08-24 10:02:10 3970 3
原创 随机采样池化--S3Pool: Pooling with Stochastic Spatial Sampling
S3Pool: Pooling with Stochastic Spatial Sampling CVPR2017 https://github.com/Shuangfei/s3pool本文将常规池化看作两个步骤: 1)以步长为1在特征图上滑动池化窗口,尺寸大小基本保持不变, leaves the spatial resolution intact 2)以一种 uniform 和 deter
2017-08-23 15:07:14 6390
原创 有效感受野--Understanding the Effective Receptive Field in Deep Convolutional Neural Networks
Understanding the Effective Receptive Field in Deep Convolutional Neural Networks NIPS 2016本文主要分析了 CNN 网络中的 Receptive Field,发现实际有效的感受野 和 理论上的感受野 差距比较大,实际有效的感受野是一个高斯分布。We introduce the notion of an eff
2017-08-23 09:12:50 8459 6
原创 人脸检测--Recurrent Scale Approximation for Object Detection in CNN
Recurrent Scale Approximation for Object Detection in CNN ICCV2017 https://github.com/sciencefans/RSA-for-object-detection本文还是针对人脸检测 中的 尺度问题 进展展开的。主要内容有以下三点: 1)首先使用一个 scale-forecast 网络来进行图像中人脸尺度的预测,
2017-08-22 15:02:29 3856
原创 人脸检测--S3FD: Single Shot Scale-invariant Face Detector
S3FD: Single Shot Scale-invariant Face Detector CVPR2017 Caffe code will be available本文针对基于 anchor 的检测器对 小的人脸检测率低的问题进行了分析和改进。基于 anchor 的目标检测发展迅速,人脸检测也有很大进展,但是对于小的人脸检测效果仍然不是很好。 the performance of an
2017-08-22 11:07:25 5779 16
原创 人脸检测--Scale-Aware Face Detection
Scale-Aware Face Detection CVPR2017针对人脸检测中的人脸多尺度问题,本文首先用一个 Scale Proposal Network (SPN) 估计出图像中人脸的尺度分布,然后按照该尺度归一化图像,再进行人脸检测 均匀分布的多尺度人脸检测有时是一种浪费,因为图像中可能只有 1-2 个尺度的人脸 所以本文的思路是先估计出人脸的尺度,有了尺度信息就可以根据该尺
2017-08-21 11:06:29 3134 2
原创 人脸识别--SphereFace: Deep Hypersphere Embedding for Face Recognition
SphereFace: Deep Hypersphere Embedding for Face Recognition CVPR2017 https://github.com/wy1iu/sphereface针对人脸识别问题,当前的损失函数基本都基于 Euclidean margin ,这里我们提出基于 angular margin 的 angular softmax (A-Softmax) l
2017-08-18 10:51:49 7092 5
原创 语义分割--Fully Convolutional DenseNets for Semantic Segmentation
The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation CVPRW 2017 Code: https://github.com/SimJeg/FC-DenseNet本文将 DenseNets 按照 FCN 方式用于 语义分割首先看看 DenseNets 该网络主要由 d
2017-08-17 16:18:49 7648
原创 语义分割--Full-Resolution Residual Networks for Semantic Segmentation in Street Scenes
Full-Resolution Residual Networks for Semantic Segmentation in Street Scenes CVPR2017 Theano/Lasagne code:https://github.com/TobyPDE/FRRN针对语义分割问题,本文侧重于分割中的物体边界精度, precise boundary adherence 通过两个 str
2017-08-17 10:22:49 3138
原创 目标检测--Enhancement of SSD by concatenating feature maps for object detection
Enhancement of SSD by concatenating feature maps for object detection BMVC 2017 本文是对SSD 的改进,通过牺牲一点速度来提高精度,主要解决SSD 两个问题:1)同一目标多次检测,2)小目标检测率不高 改进的地方: 1)将不同尺度的 特征图 融合起来; 2)增加 feature pyramid 网络层的特征图数量
2017-08-16 16:41:38 2285
原创 目标检测--DSOD: Learning Deeply Supervised Object Detectors from Scratch
DSOD: Learning Deeply Supervised Object Detectors from Scratch ICCV2017 https://github.com/szq0214/DSOD针对目标检测问题,本文提出了不需要预训练模型的检测算法,可以看作 SSD + DenseNet 的结合以前的目标检测算法基本都是先在 ImageNet 数据库上进行预训练,然后再微调。这个微
2017-08-16 14:48:01 11124
原创 目标检测--SSD: Single Shot MultiBox Detector
SSD: Single Shot MultiBox Detector ECCV2016 https://github.com/weiliu89/caffe/tree/ssd 针对目标检测问题,本文取消了候选区域提取步骤,通过采用一系列设计在检测上得到较好的精度和速度contributions : 1)提出的 SSD 比 YOLO v1 速度快,精度好, 和 Faster R-CNN 精度差不
2017-08-15 16:06:31 3226
原创 目标检测--Focal Loss for Dense Object Detection
Focal Loss for Dense Object Detection ICCV2017 https://arxiv.org/abs/1708.02002本文算是用简单的方法解决复杂的问题了,好的思想是简单的针对目标检测,目前有两类主流算法: two-stage detectors 和 one-stage detectors, two-stage detectors 的精度好但是速度慢,on
2017-08-15 10:03:04 2334
原创 视频目标检测--Flow-Guided Feature Aggregation for Video Object Detection
Flow-Guided Feature Aggregation for Video Object Detection https://arxiv.org/abs/1703.10025 Our framework is principled, and on par with the best engineered systems winning the ImageNet VID challenge
2017-08-04 14:51:21 5011
原创 视频检测分割--Deep Feature Flow for Video Recognition
Deep Feature Flow for Video Recognition CVPR2017 Code: https://github.com/msracver/Deep-Feature-Flow基于单帧的目标检测和分割已经做的比较成熟,但是基于视频的目标检测和分割目前还是有问题的,最主要的问题就是直接将单帧的算法用于视频,计算量比较大,做不到实时。这里我们只对关键帧计算CNN特征提取,然后
2017-08-04 10:23:20 7325 2
原创 有方向的CNN--Oriented Response Networks
Oriented Response Networks CVPR2017 http://yzhou.work/ORN/ Code: https://github.com/ZhouYanzhao/ORN以前的CNN主要通过 pool 来得到小的角度变化不变性,本文设计了一个CNN 可以得到目标的角度信息,也可以做到完全的旋转不变性。 简单的来说,对以前学习到的滤波器 f 旋转 N个角度(例如,每
2017-08-03 16:58:08 3903
原创 语义分割--Efficient and Robust Deep Networks for Semantic Segmentation
Efficient and Robust Deep Networks for Semantic SegmentationCode: https://lmb.informatik.uni-freiburg.de/Publications/2017/OB17a/在 Up-Convolutional Networks 的基础上,提出了 Part-Net,Fast-Net,M-Net。 三个网络的侧重点不
2017-08-02 15:27:28 2921
原创 语义分割--Efficient Deep Models for Monocular Road Segmentation
Efficient Deep Models for Monocular Road Segmentation code: https://lmb.informatik.uni-freiburg.de/Publications/2016/OB16b/针对路面检测和分割问题,本文结合FCN 和 U-Net 提出一个网络,在速度和精度方面得到不错的效果 KITTI benchmark lane/ro
2017-08-02 10:50:19 3049 1
原创 查看 CUDA cudnn 版本 & 测试 cuda 和 cudnn 有效性
https://medium.com/@changrongko/nv-how-to-check-cuda-and-cudnn-version-e05aa21daf6ccuda 版本 cat /usr/local/cuda/version.txtcudnn 版本 cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
2017-08-01 15:27:09 244438 9
Accuracy of Laplacian Edge Detectors
2011-10-12
The Canny Edge Detector Revisited
2011-08-11
OpenCV 2 Computer Vision Application Programming Cookbook
2011-06-24
Learning based Symmetric Features Selection for Vehicle Detection
2011-04-11
Intensity and Edge-Based Symmetry Detection Applied to Car-Following
2011-04-11
Accurate Robust Symmetry Estimation
2011-04-11
Approach of vehicle segmentation based on texture character
2011-04-01
Method of removing moving shadow based on texture
2011-04-01
Environmentally Robust Motion Detection for Video Surveillance
2011-03-17
Optimal multi-level thresholding using a two-stage Otsu optimization approach
2011-03-17
A Background Reconstruction Method Based on Double-background
2011-03-17
Statistical Change Detection by the Pool Adjacent Violators Algorithm
2011-03-17
Cooperative Fusion of Stereo and Motion
2011-03-09
A Treatise on Mathematical Theory of Elasticity (1944)(ISBN 0486601749)
2011-02-27
A Treatise on Mathematical Theory of Elasticity (1944)(ISBN 0486601749)
2011-02-27
Love, A Treatise on Mathematical Theory of Elasticity (1944)(ISBN 0486601749)
2011-02-27
Computation of Real-Time Optical Flow Based on Corner Features
2011-02-24
II-LK – A Real-Time Implementation for Sparse Optical Flow
2011-02-24
Medical Image Reconstruction A Conceptual Tutorial --pdf
2011-02-24
Extraction and recognition of license plates of motorcycles and vehicles on highways
2011-02-22
High Performance Implementation of License Plate Recognition in Image Sequences
2011-02-22
Vs-star-- A visual interpretation system for visual surveillance
2011-02-22
Robust fragments-based tracking with adaptive feature selection
2011-02-22
Robust and automated unimodal histogram thresholding and potential applications
2011-02-22
角点检测方法研究-- 毛雁明, 兰美辉
2011-02-22
Corner Detection Algorithms for Digital Images in Last Three Decades
2011-02-22
Application of Shape Analysis Techniques for the Classification of Vehicles
2011-02-22
Solving the process of hysteresis without determining the optimal thresholds
2011-02-22
百面机器学习.pdf
2019-06-01
CLIP-Q CVPR2018 code
2018-10-30
Vehicle Detection and Tracking in Car Video Based on Motion Model
2011-10-15
Projection and Least Square Fitting
2011-10-15
A Review of Computer Vision Techniques for the Analysis of Urban Traffic
2011-10-15
On Improving the Efficiency of Tensor Voting
2011-10-11
Fast LOG Filtering Using Recursive Filters
2011-10-11
A discrete expression of Canny's criteria for step
2011-10-11
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人