• 显著性目标检测代码
2022-05-18 10:40:10

网上很多资源画PR曲线的代码并不能用，这里放一个python版的评估代码
https://github.com/lartpang/PySODEvalToolkit

更多相关内容
• 显著性目标检测代码全汇总！(包含2D、3D、4D以及Video) 转载于极市开发者原创投稿https://mp.weixin.qq.com/s/ZSytEsVSjBU_zjy5YPxPeA SOD CNNs-based Read List In this repository, we mainly focus on deep ...

# 显著性目标检测代码全汇总！(包含2D、3D、4D以及Video)

## 转载于极市开发者原创投稿https://mp.weixin.qq.com/s/ZSytEsVSjBU_zjy5YPxPeA

这里是加以翻译

在这个知识库中，我们主要关注基于深度学习的显著性方法(2D RGB, 3D RGB-D, Video SOD and 4D Light Field) ，并提供总结(Code and Paper).。我们希望这份回购能帮助大家更好地理解深度学习时代的显著性检测。 评估指标。

2D SOD: Add five papesr ECCV20
3D SOD: Add nine papers ECCV20 and two ACMM20 papers
Video SOD : Add three papers ECCV20, Continuously Updating!

# 2D RGB Saliency Detection

## 2020

01AAAIProgressive Feature Polishing Network for Salient Object DetectionPaper/Code
02AAAIGlobal Context-Aware Progressive Aggregation Network for Salient Object DetectionPaper/Code
03AAAIF3Net: Fusion, Feedback and Focus for Salient Object DetectionPaper/Code
05AAAIMulti-Type Self-Attention Guided Degraded Saliency DetectionPaper/Code
06CVPRWeakly-Supervised Salient Object Detection via Scribble AnnotationsPaper/Code
07CVPRTaking a Deeper Look at the Co-salient Object DetectionPaper/Code
08CVPRMulti-scale Interactive Network for Salient Object DetectionPaper/Code
09CVPRInteractive Two-Stream Decoder for Accurate and Fast Saliency DetectionPaper/Code
10CVPRLabel Decoupling Framework for Salient Object DetectionPaper/Code
11CVPRAdaptive Graph Convolutional Network with Attention Graph Clustering for Co-saliency DetectionPaper/Code
🚩 12ECCVHighly Efficient Salient Object Detection with 100K ParametersPaper/Code
🚩 13ECCVn-Reference Transfer Learning for Saliency PredictionPaper/Code
🚩 13ECCVLearning Noise-Aware Encoder-Decoder from Noisy Labels by Alternating Back-Propagation for Saliency DetectionPaper/Code
🚩 15ECCVSuppress and Balance: A Simple Gated Network for Salient Object DetectionPaper/Code

## 2019

01CVPRAFNet: Attentive Feedback Network for Boundary-aware Salient Object DetectionPaper/Code
02CVPRBASNet: Boundary Aware Salient Object DetectionPaper/Code
03CVPRCPD: Cascaded Partial Decoder for Accurate and Fast Salient Object DetectionPaper/Code
04CVPRMulti-source weak supervision for saliency detectionPaper/Code
05CVPRMLMSNet:A Mutual Learning Method for Salient Object Detection with intertwined Multi-SupervisionPaper/Code
06CVPRCapSal: Leveraging Captioning to Boost Semantics for Salient Object DetectionPaper/Code
07CVPRPoolNet: A Simple Pooling-Based Design for Real-Time Salient Object DetectionPaper/Code
08CVPRAn Iterative and Cooperative Top-down and Bottom-up Inference Network for Salient Object DetectionPaper/Code
09CVPRPyramid Feature Attention Network for Saliency detectionPaper/Code
10AAAIDeep Embedding Features for Salient Object DetectionPaper/Code
11ICIPSalient Object Detection Via Deep Hierarchical Context Aggregation And Multi-Layer SupervisionPaper/Code
12IEEE TCSVTAADF-Net: Aggregating Attentional Dilated Features for Salient ObjectPaper/Code
13IEEE TCybROSA: Robust Salient Object Detection against Adversarial AttacksPaper/Code
14arXivDSAL-GAN: DENOISING BASED SALIENCY PREDICTION WITH GENERATIVE ADVERSARIAL NETWORKSPaper/Code
15arXivSAC-Net: Spatial Attenuation Context for Salient Object DetectionPaper/Code
16arXivSE2Net: Siamese Edge-Enhancement Network for Salient Object DetectionPaper/Code
17arXivRegion Refinement Network for Salient Object DetectionPaper/Code
18arXivContour Loss: Boundary-Aware Learning for Salient Object SegmentationPaper/Code
19arXivOGNet: Salient Object Detection with Output-guided Attention ModulePaper/Code
20arXivEdge-guided Non-local Fully Convolutional Network for Salient Object DetectionPaper/Code
21ICCVFLoss:Optimizing the F-measure for Threshold-free Salient Object DetectionPaper/Code
22ICCVStacked Cross Refinement Network for Salient Object DetectionPaper/Code
23ICCVSelectivity or Invariance: Boundary-aware Salient Object DetectionPaper/Code
24ICCVHRSOD:Towards High-Resolution Salient Object DetectionPaper/Code
25ICCVEGNet:Edge Guidance Network for Salient Object DetectionPaper/Code
26ICCVStructured Modeling of Joint Deep Feature and Prediction Refinement for Salient Object DetectionPaper/Code
27ICCVEmploying Deep Part-Object Relationships for Salient Object DetectionPaper/Code
28NeurIPSDeep Robust Unsupervised Saliency Prediction With Self-SupervisionPaper/Code

## 2018

01CVPRA Bi-Directional Message Passing Model for Salient Object DetectionPaper/Code
02CVPRPiCANet: Learning Pixel-wise Contextual Attention for Saliency DetectionPaper/Code
03CVPRPAGR: Progressive Attention Guided Recurrent Network for Salient Object DetectionPaper/Code
04CVPRLearning to promote saliency detectorsPaper/Code
05CVPRDetect Globally, Refine Locally: A Novel Approach to Saliency DetectionPaper/Code
06CVPRSalient Object Detection Driven by Fixation PredictionPaper/Code
07IJCAIR3Net: Recurrent Residual Refinement Network for Saliency DetectionPaper/Code
08IJCAILFR: Salient Object Detection by Lossless Feature ReflectionPaper/Code
09ECCVContour Knowledge Transfer for Salient Object DetectionPaper/Code
10ECCVReverse Attention for Salient Object DetectionPaper/Code
11IEEE TIPAn unsupervised game-theoretic approach to saliency detectionPaper/Code
12arXivAgile Amulet: Real-Time Salient Object Detection with Contextual AttentionPaper/Code
13arXivHyperFusion-Net: Densely Reflective Fusion for Salient Object DetectionPaper/Code
14arXiv(TBOS)Three Birds One Stone: A Unified Framework for Salient Object Segmentation, Edge Detection and Skeleton ExtractionPaper/Code

## 2017

01CVPRDSS: Deeply Supervised Salient Object Detection with Short ConnectionsPaper/Code
02CVPRNon-Local Deep Features for Salient Object DetectionPaper/Code
03CVPRLearning to Detect Salient Objects with Image-level SupervisionPaper/Code
04CVPRSalGAN: visual saliency prediction with adversarial networksPaper/Code
05ICCVA Stagewise Refinement Model for Detecting Salient Objects in ImagesPaper/Code
06ICCVAmulet: Aggregating Multi-level Convolutional Features for Salient Object DetectionPaper/Code
07ICCVLearning Uncertain Convolutional Features for Accurate Saliency DetectionPaper/Code

## 2016

01CVPRDHSNet: Deep hierarchical saliency network for salient object detectionPaper/Code
02CVPRELD: Deep Saliency with Encoded Low level Distance Map and High Level FeaturesPaper/Code
03ECCVRFCN: Saliency detection with recurrent fully convolutional networksPaper/Code

# 3D RGB-D Saliency Detection

## 2020

01IEEE TIPICNet: Information Conversion Network for RGB-D Based Salient Object DetectionPaper/Code
02CVPRJL-DCF: Joint Learning and Densely-Cooperative Fusion Framework for RGB-D Salient Object DetectionPaper/Code
03CVPRUC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational AutoencodersPaper/Code
04CVPRA2dele: Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object DetectionPaper/Code
05CVPRSelect, Supplement and Focus for RGB-D Saliency DetectionPaper/Code
06CVPRLearning Selective Self-Mutual Attention for RGB-D Saliency DetectionPaper/Code
🚩 07ECCVAccurate RGB-D Salient Object Detection via Collaborative LearningPaper/Code
🚩 08ECCVCross-Modal Weighting Network for RGB-D Salient Object DetectionPaper/Code
🚩 09ECCVBBS-Net: RGB-D Salient Object Detection with a Bifurcated Backbone Strategy NetworkPaper/Code
🚩 10ECCVHierarchical Dynamic Filtering Network for RGB-D Salient Object DetectionPaper/Code
🚩 11ECCVProgressively Guided Alternate Refinement Network for RGB-D Salient Object DetectionPaper/Code
🚩 12ECCVRGB-D Salient Object Detection with Cross-Modality Modulation and SelectionPaper/Code
🚩 13ECCVCascade Graph Neural Networks for RGB-D Salient Object DetectionPaper/Code
🚩 14ECCVA Single Stream Network for Robust and Real-time RGB-D Salient Object DetectionPaper/Code
🚩 15ECCVAsymmetric Two-Stream Architecture for Accurate RGB-D Saliency DetectionPaper/Code
🚩 16ACMMIs Depth Really Necessary for Salient Object Detection?Paper/Code
🚩 17ACMMA Top-down and Adaptive Fusion Network for RGB-D Salient Object DetectionPaper/Code

## 2019

01ICCVDMRA: Depth-induced Multi-scale Recurrent Attention Network for Saliency DetectionPaper/Code
02CVPRCPFP: Contrast Prior and Fluid Pyramid Integration for RGBD Salient Object DetectionPaper/Code
03IEEE TIPThree-stream Attention-aware Network for RGB-D Salient Object DetectionPaper/Code
04IEEE PRMulti-modal fusion network with multi-scale multi-path and cross-modal interactions for RGB-D salient object detectionPaper/Code
05arXivAFNet: Adaptive Fusion for RGB-D Salient Object DetectionPaper/Code
06IEEE TNNLSD3Net:Rethinking RGB-D Salient Object Detection: Models, Datasets, and Large-Scale BenchmarksPaper/Code
07arXivCNN-based RGB-D Salient Object Detection: Learn, Select and FusePaper/Code

## 2018

01CVPRPCA: Progressively Complementarity-aware Fusion Network for RGB-D Salient Object DetectionPaper/Code
02IEEE TIPCo-saliency detection for RGBD images based on multi-constraint feature matching and cross label propagationPaper/Code
03ICMEPDNet: Prior-Model Guided Depth-enhanced Network for Salient Object DetectionPaper/Code

## 2017

01ICCVLearning RGB-D Salient Object Detection using background enclosure, depth contrast, and top-down featuresPaper/Code
02IEEE TIPDF: RGBD Salient Object Detection via Deep FusionPaper/Code
03IEEE TCybCTMF: Cnns-based rgb-d saliency detection via cross-view transfer and multiview fusionPaper/Code

01MTARGBD co-saliency detection via multiple kernel boosting and fusionPaper/Code
02ICCV17An Innovative Salient Object Detection Using Center-Dark Channel PriorPaper/Code
03IEEE SPLSaliency detection for stereoscopic images based on depth confidence analysis and multiple cues fusionPaper/Code
04IEEE SPLRGBD Co-saliency Detection via Bagging-Based ClusteringPaper/Code
05CVPRExploiting Global Priors for RGB-D Saliency DetectionPaper/Code

# 4D Light Field Saliency Detection

01TOMMMCA: Saliency Detection on Light Field: A Multi-Cue ApproachPaper/Code
02IJCAIDILF: Saliency Detection with a Deeper Investigation of Light FieldPaper/Code
03CVPRWSC: A Weighted Sparse Coding Framework for Saliency DetectionPaper/Code
04IEEE PAMISaliency Detection on Light-FieldPaper/Code
05ICCVDeep Learning for Light Field Saliency DetectionPaper/Code
06NeurIPSMemory-oriented Decoder for Light Field Salient Object DetectionPaper/Code
07AAAIExploit and Replace: An Asymmetrical Two-Stream Architecture for Versatile Light Field Saliency DetectionPaper/Code

# Video Salient Object Detection

## 2020

01CVPRSTAViS: Spatio-Temporal AudioVisual Saliency NetworkPaper/Code
🚩 02ECCVUnified Image and Video Saliency ModelingPaper/Code
🚩 03ECCVMeasuring the importance of temporal features in video saliencyPaper/Code
🚩 04ECCVTENet: Triple Excitation Network for Video Salient Object DetectionPaper/Code

## 2019

01ICCVMotion Guided Attention for Video Salient Object DetectionPaper/Code
02ICCVSemi-Supervised Video Salient Object Detection Using Pseudo-LabelsPaper/Code
03ICCVTemporally-Aggregating Spatial Encoder-Decoder Network for Video Saliency DetectionPaper/Code
04ICCVRANet：Ranking attention Network for Fast Video Object SegmentationPaper/Code
05CVPRShifting More Attention to Video Salient Objection DetectionPaper/Code
06CVPRLearning Unsupervised Video Object Segmentation through Visual AttentionPaper/Code
07CVPRSee More, Know More: Unsupervised Video Object Segmentation with Co-Attention Siamese NetworksPaper/Code

## 2018

01ECCVPyramid Dilated Deeper CoonvLSTM for Video Salient Object DetectionPaper/Code
02ECCVDeepVS: A Deep Learning Based Video Saliency Prediction ApproachPaper/Code
03CVPRRevisiting Video Saliency: A Large-scale Benchmark and a New ModelPaper/Code
04CVPRFlow Guided Recurrent Neural Encoder for Video Salient Object DetectionPaper/Code
05IEEE TIPVideo Salient Object Detection via Fully Convolutional NetworksPaper/Code

## 2017

01IEEE TIPLearning to Detect Video Saliency with HEVC FeaturesPaper/Code

# Earlier Methods

01IEEE TIP15Salient object detection: A benchmarkPaper/Code
02IEEE TCSVT18Review of visual saliency detectionwith comprehensive informationPaper/Code
03ACM TIST18A review of co-saliency detection algorithms: Fundamentals, applications, and challengesPaper/Code
04IEEE TSP18Advanced deep-learning techniques for salient and category-specific object detection: A surveyPaper/Code
05IJCV18Attentive systems: A surveyPaper/Project
06ECCV18Salient Objects in Clutter: Bringing Salient Object Detection to the ForegroundPaper/Code
07CVM18Salient object detection: A surveyPaper/Code
08IEEE TNNLS19Salient Object detection with deep learning: AreviewPaper/Code
09arXiv19Salient Object Detection in the Deep Learning Era-An In-Depth SurveyPaper/Code
10GitHub20RGB-D Salient Object Detection: A SurveyPaper/Code

The part of the collection is thanks to Deng-Ping Fan and Tao Zhou.

• Salient Object Detection in the Deep Learning Era: An In-Depth Survey. paper link.
• RGB-D Salient Object Detection: A Survey. project link.

# 与最先进的比较

• Here 包括几乎所有2D显著目标检测算法的性能比较.
• Here i包括几乎所有3D RGB-D显著目标检测算法的性能比较。

# Evaluation Metrics【评价指标】

• Saliency maps evaluation.【特征图评价】
该链接几乎包括用于显著对象检测的所有评估指标，包括E-Measure、S-Measure、F-Measure、MAE Score和PR曲线或BAR指标。 你可以在这里找到 【链接】.

• Saliency Dataset evaluation.【显著性数据集评估】
该方法可以在二进制显著性数据集上计算Obj.Area和Obj.Contrast的比值。此工具箱包含两个评估指标，包括obj(Object).Area和obj.Contrast。
你可以在这里找到【链接】.

### AI会议截止日期

展开全文
• 显著性目标检测对比实验评估代码

放一个显著性目标检测评估代码matlab版链接
https://github.com/weijun88/F3Net

展开全文
• 论文仅供学习参考使用。...提出使用目标性作为先验信息得到前景显著图，并且利用乘法运算将其与基于背景先验信息计算的显著图相融合，然后进行空间优化得到单尺度下的显著图，最终显著图为多尺度显著图的加权融合．
• 显著性目标检测是机器视觉领域的研究热点,具有广泛的应用前景。针对现有显著性目标检测算法存在的显著区域检测不均匀、边缘表示模糊等问题,提出一种双注意力循环卷积显著性目标检测算法。在U-Net全卷积骨干网络中...
• 小型红外目标检测的绝对方向平均差（ADMD）算法 以下论文的MATLAB和OpenCV两种实现：使用绝对方向均值差算法的快速而强大的小型红外目标检测 如果您在研究中使用这些代码，请引用以下论文： 的MATLAB 您可以在MATLAB...
• 图像融合是一种重要的增强图像信息的技术方法,如何对同一目标的多源遥感图像数据进行有效的融合,最大限度地利用多源遥感数据中的有用信息,提高系统的正确识别、判断和决策能力,这是遥感数据融合研究的重要内容之一。...

## 1 简介

图像融合是一种重要的增强图像信息的技术方法,如何对同一目标的多源遥感图像数据进行有效的融合,最大限度地利用多源遥感数据中的有用信息,提高系统的正确识别、判断和决策能力,这是遥感数据融合研究的重要内容之一。图像融合技术的发展经历了3 个阶段:(1)简单的图像融合方法,如RGB假彩色合成、HIS彩色变换、PCA主分量变换法等; (2)随着塔式算子的提出,在融合领域也出现了一些较为复杂的模型;(3)用小波变换的多尺度分析替代塔式算法。传统的图像数据融合方法对中、高分辨率的遥感图像的数据融合一般都能取得比较理想的效果,但对于低分辨率的遥感图像数据融合效果并不明显。具有“数学显微镜”之称的小波变换同时在时域和频域具有分辨率,对高频分量采用逐渐精细的时域或空域步长,可以聚焦到分析对象的任意细节,对于剧烈变化的边缘,比常规的傅里叶变换具有更好的适应性。由于小波变换具有的特点,使它很快在图像处理中得到广泛的应用。与传统的数据融合方法相比,小波融合方法不仅能够针对输入图像的不同特征来合理选择小波基以及小波变换的次数,而且在融合操作时又可以根据实际需要来引入双方的细节信息。从而表现出更强的针对性和实用性,融合效果更好。另外,从实施过程的灵活性方面评价,HIS彩色变换只能而且必须同时对三个波段进行融合操作,PCA主分量变换法的输入图像必须有三个或三个以上,而小波方法则能够完成对单一波段或多个波段的融合运算,对于单个黑白图像的融合,小波方法更是唯一的选择。本文提出了一种基于小波变换的融合方法,使得融合图像在最大限度保留多波段光谱信息的同时,提高了清晰度和空间分辨率。并在MATLAB环境下对该方法进行了实例分析,从图像清晰度、信息墒、信噪比等几个方面对结果做了深入的分析与对比,发现融合后的图像均值和方差基本保持不变,图像信噪比为20db左右,说明融合后的图像基本保持了原始图像的光谱特性,而信息熵和清晰度有明显的提高。因此基于小波变换的Mallat多分辨率分析可有效地用于低分辨率多光谱遥感图像的数据融合,融合后的图像在信息含量、细节、目标解析水平等方面明显优于原图像。

## 2 部分代码

function [R, G, B] = Lab2RGB(L, a, b)%LAB2RGB Convert an image from CIELAB to RGB%% function [R, G, B] = Lab2RGB(L, a, b)% function [R, G, B] = Lab2RGB(I)% function I = Lab2RGB(...)%% Lab2RGB takes L, a, and b double matrices, or an M x N x 3 double% image, and returns an image in the RGB color space.  Values for L are in% the range [0,100] while a* and b* are roughly in the range [-110,110].% If 3 outputs are specified, the values will be returned as doubles in the% range [0,1], otherwise the values will be uint8s in the range [0,255].%% This transform is based on ITU-R Recommendation BT.709 using the D65% white point reference. The error in transforming RGB -> Lab -> RGB is% approximately 10^-5.  %% See also RGB2LAB. ​% By Mark Ruzon from C code by Yossi Rubner, 23 September 1997.% Updated for MATLAB 5 28 January 1998.% Fixed a bug in conversion back to uint8 9 September 1999.% Updated for MATLAB 7 30 March 2009.​if nargin == 1  b = L(:,:,3);  a = L(:,:,2);  L = L(:,:,1);end​% ThresholdsT1 = 0.008856;T2 = 0.206893;​[M, N] = size(L);s = M * N;L = reshape(L, 1, s);a = reshape(a, 1, s);b = reshape(b, 1, s);​% Compute YfY = ((L + 16) / 116) .^ 3;YT = fY > T1;fY = (~YT) .* (L / 903.3) + YT .* fY;Y = fY;​% Alter fY slightly for further calculationsfY = YT .* (fY .^ (1/3)) + (~YT) .* (7.787 .* fY + 16/116);​% Compute XfX = a / 500 + fY;XT = fX > T2;X = (XT .* (fX .^ 3) + (~XT) .* ((fX - 16/116) / 7.787));​% Compute ZfZ = fY - b / 200;ZT = fZ > T2;Z = (ZT .* (fZ .^ 3) + (~ZT) .* ((fZ - 16/116) / 7.787));​% Normalize for D65 white pointX = X * 0.950456;Z = Z * 1.088754;​% XYZ to RGBMAT = [ 3.240479 -1.537150 -0.498535;       -0.969256  1.875992  0.041556;        0.055648 -0.204043  1.057311];​RGB = max(min(MAT * [X; Y; Z], 1), 0);​R = reshape(RGB(1,:), M, N);G = reshape(RGB(2,:), M, N);B = reshape(RGB(3,:), M, N); ​if nargout < 2  R = uint8(round(cat(3,R,G,B) * 255));end​​

## 4 参考文献

[1]周国庆. 基于视觉显著性的图像目标检测设计与实现[D]. 西安电子科技大学.

### 博主简介：擅长智能优化算法、神经网络预测、信号处理、元胞自动机、图像处理、路径规划、无人机等多种领域的Matlab仿真，相关matlab代码问题可私信交流。

部分理论引用网络文献，若有侵权联系博主删除。

展开全文
• 提出了一种利用稀疏表达检测多幅图像中协同显著目标的方法。首先用独立变量分析方法训练得到自然图像一组稀疏基,接着求出检测图像的...实验结果表明,该方法正确有效,具有和人类视觉特性相符合的显著性目标检测效果。
• Learning to Promote Saliency Detectors论文阅读 ...相反，将 DNN 拟合为一个嵌入函数，以将像素和显著/背景区域的属性映射到度量空间。显着/背景区域的属性被映射为度量空间中的锚点。然后，在该空间中
• 针对视频序列中的显著性运动目标检测问题，提出一种基于图像频域中的多尺度多特征显著性运动目标的快速检测方法．所提算法在对视频序列多尺度运算的基础上，通过提取视频序列底层特征，分析其离散余弦频域中的时空域...
• 显著性目标检测 显著性目标检测旨在从输入图像上识别出最引人注目的对象，简单来说，这个研究方向希望能够识别出图像的主体。 以深度学习方法进行显著性目标检测 虽然手制特征允许传统显著性目标检测方法实时进行...
• 显著性检测中画PR曲线的代码，需要自己生成的显著图和Ground Truth。
• 在图像分割、目标检测、场景感知等许多图像处理任务中,图像中不同区域对视觉系统刺激程度不同引起的视觉显著性信息将系统资源优先集中于感兴趣区域进行计算分析,降低了处理过程的复杂性,为后续处理提供了极大的便
• 论文：RGB-D Salient Object Detection: A Survey 论文下载：RGB-D Salient Object Detection: A Survey 代码：https://github.com/taozh2017/RGBD-SODsurvey ...显著性目标检测（SOD）可模拟人类视觉感
• 针对具有杂乱背景图像的显著目标检测问题，提出了一种无需任何先验知识，通过分析计算区域平均显著值的对比度来提取显著目标的方法．根据显著图，计算出显著目标的最小边界框与其周围区域的显著性差异，且通过折半...
• 点击上方“AI算法修炼营”，选择“星标”公众号精选作品，第一时间送达本文是收录于CVPR2020的有关显著性目标检测的文章，主要的创新点在特征聚合操作，可以迁移到其他需要融合深层和浅层特...
• 该存储库包含ICIP论文《高光谱图像中的显着目标检测》中描述的算法的源代码。 可以在本文中找到更多详细信息。 该软件包已在64位Windows计算机上使用Matlab 2013a进行了测试。 此代码仅用于研究目的。 引用 如果您...
• 现有的方法大多依靠RGB信息来区分显著性目标，在一些复杂的场景中存在困难。为了解决这一问题，近年来许多基于RGB的网络被提出，它们采用深度图作为独立的输入，并将特征与RGB信息融合。借鉴RGB方法和RGBD方法的优点...
• 里面有一部分CRF后处理预测图的代码，这里我就直接扒过来了。 使用方法： if args['crf_refine']: prediction = crf_refine(np.array(img), np.array(prediction)) 源码： import pydensecrf.densecrf as dcrf ...
• 显著性目标检测模型通常需要花费大量的计算成本才能对每个像素进行精确的预测，因此这使得其几乎不适用于低功耗的设备。 本文旨在通过提高网络计算效率来缓解计算花费与模型性能之间的矛盾。本文提出了一种灵活的...
• 显著性目标检测模型评价指标 之 PR曲线原理与实现代码 目录 显著性目标检测模型评价指标 之 PR曲线原理与实现代码 目录 一、PR曲线原理 计算方法 阈值选取 二、Matlab代码 著作权归作者所有。商业转载请...
• 图像显著性区域检测技术可以实现图像显著性信息的提取.可靠快速的显著性检测能够为内容感知的图像编辑,图像分割,图像检索等应用提供有价值的参考信息;能够缓解图像内容理解与图像底层特征之间的隔阂,使得更高层的...
• Deep-learning based salient（显著的）object detection methods achieve great progress.However, the variable scale and unknown category of salient objects are great challenges all the time.These are ...
• FT 显著性检测算法从频域出发，利用高斯低通滤波，计算 CIE Lab 颜色空间中单个像素和图像所有像素平均值的欧氏距离作为该像素的显著值。FT 算法在强调显著对象的同时，能够建立较为清晰的边界，对纹理和噪声产生的...
• 显著性目标检测近年来论文和代码 包括2020CVPR，2021CVPR，2019ICCV，2020ECCV 链接：https://pan.baidu.com/s/14hMQcSjdXkEHZQvKy6HqIw 提取码：zfka

...