RNN

Long Short- term Memory(LSTM1997)Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation(GRU2014)Depth-Gated Recurrent Neural Networks 2015A Clockwork RNN 2014Grid Long Short-Term Memory 2015 6

difference of various RNN units

LSTM: A Search Space Odyssey 2015An Empirical Exploration of Recurrent Network Architecture 2015 3.

ATTENTION

Recurrent model for visual attentionMultiple Object Recognition With Visual AttentionShow, Attend and Tell: Neural Image Caption Generation with Visual AttentionDeep Recurrent Visual Attention Model 4.

RNN MODELS

DRAW: A Recurrent Neural Network For Image GenerationLEARNING STOCHASTIC RECURRENT NETWORKSBidirectional Recurrent Neural Networks

GOOD

VISUALIZING AND UNDERSTANDING RECURRENT NETWORKSAn Empirical Exploration of Recurrent Network ArchitecturesRECURRENT NEURAL NETWORK REGULARIZATIONSequence to Sequence Learning with Neural Networks

Application

Image classification with recurrent attention models

TORCH

recurrent model of visual attention torch 官方解释

FV 和CNN的优劣 1. Analyzing Classifiers: Fisher Vectors and Deep Neural Networks

CVPR2016

http://www.cv-foundation.org/openaccess/CVPR2016.py

准备Fisher vector 看的论文

Multi-Scale Orderless Pooling of Deep Convolutional Activation FeaturesAggregating local descriptors into a compact image representationLeveraging Structure from Motion to Learn Discriminative Codebooks for Scalable Landmark ClassificationImproving the Fisher Kernel for Large-Scale Image Classification Hybrid multi-layer deep CNN/aggregator feature for image classificationFisher vectors meet Neural Networks: A hybrid classification architecture Image Classification with CNN-based Fisher Vector Coding 这篇文章细节太少，不适合作参考，可以适度阅读，领会精神Fully Convolutional Networks for Semantic Segmentation fully convolutional networks 的框架就是在这里看懂的，看懂之后觉得很简单，但是有些论文就是连这些最简单的东西都说不清楚。牛逼的论文就是能够让小白也能看懂，看不懂的可以引出出处，进而进一步搜索。 TPAMI 2016

分类器

Do we Need Hundreds of Classifiers to Solve Real World Classification Problems? 该文对100多种分类器进行了实验对比，最后证明并行随机森林能够取得最好的分类效果。SVM_c和随机森林算是结果最好的两组family。

frequency domain

Advanced Image Classification using Wavelets and Convolutional Neural Networks Frequency and Space Domain Features for Image Classification Using Gaussian Mixture Models Supervised Image Classification Using Deep Convolutional Wavelets NetworkCompressed-Domain Ship Detection on Spaceborne Optical Image using Deep Neural Network and Extreme Learning MachineCombining time- and frequency-domain convolution in convolutional neural network-based phone recognition

杂记

Bilinear CNN Models for Fine-grained Visual Recognitionthe Application of Two-level Attention Models in Deep Convolutional Neural Network for Fine-grained Image ClassificationModelling local deep convolutional neural network features to improve fine-grained image classificationPart-based R-CNNs for Fine-grained Category DetectionHistology Image Classification using Supervised Classification and Multimodal Fusion Automated gland and nuclei segmentation for grading of prostate and breast cancer histopathology: 使用分割腺体和细胞核来评分，服

WT Based Sonoelastography Prostate Cancer Image Classification Using Back Propagation Neural Network

CURVELET-BASED CLASSIFICATION OF PROSTATE CANCER HISTOLOGICAL IMAGES OF CRITICAL GLEASON SCORES

Deep convolutional activation features for large scale Brain Tumor histopathology image classification and segmentation 把提取到的4096维特征如何结合在一起，我参考的就是这篇文档

——–2017.04.19