MatConvNet - Convolutional Neural Networks for MATLAB

Abstract

MatConvNet is an implementation of Convolutional Neural Networks (CNNs) for MATLAB. The toolbox is designed with an emphasis on simplicity and flexibility. It exposes the building blocks of CNNs as easy-to-use MATLAB functions, providing routines for computing linear convolutions with filter banks, feature pooling, and many more. In this manner, MatConvNet allows fast prototyping of new CNN architectures; at the same time, it supports efficient computation on CPU and GPU allowing to train complex models on large datasets such as ImageNet ILSVRC. This document provides an overview of CNNs and how they are implemented in MatConvNet and gives the technical details of each computational block in the toolbox.

Updated for release v1.0-beta20

Cited by in corpus (116)

• Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
• Deep Convolutional Neural Network for Inverse Problems in Imaging
• VoxCeleb2: Deep Speaker Recognition
• Blind Image Quality Assessment Using A Deep Bilinear Convolutional Neural Network
• A Discriminatively Learned CNN Embedding for Person Re-identification
• Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computing
• Learning-Based View Synthesis for Light Field Cameras
• Particular object retrieval with integral max-pooling of CNN activations
• Machine learning in acoustics: theory and applications
• NiftyNet: a deep-learning platform for medical imaging
• Deep Label Distribution Learning with Label Ambiguity
• Convolutional Neural Network Based Metal Artifact Reduction in X-ray Computed Tomography
• Skin lesion detection based on an ensemble of deep convolutional neural network
• Deep Learning for Visual Tracking: A Comprehensive Survey
• Learning with a Wasserstein Loss
• Local Learning with Deep and Handcrafted Features for Facial Expression Recognition
• Maximum Entropy Deep Inverse Reinforcement Learning
• Context-aware Deep Feature Compression for High-speed Visual Tracking
• Rotation equivariant vector field networks
• Region Convolutional Features for Multi-Label Remote Sensing Image Retrieval
• Unsupervised Person Re-identification by Deep Asymmetric Metric Embedding
• Deep Residual Learning for Compressed Sensing CT Reconstruction via Persistent Homology Analysis
• A Theory of Generative ConvNet
• Incremental Learning in Deep Convolutional Neural Networks Using Partial Network Sharing
• Zero-Shot Learning via Semantic Similarity Embedding
• Mask-CNN: Localizing Parts and Selecting Descriptors for Fine-Grained Image Recognition
• deSpeckNet: Generalizing Deep Learning Based SAR Image Despeckling
• Attribute2Image: Conditional Image Generation from Visual Attributes
• Batch-normalized Maxout Network in Network
• Exploiting deep residual networks for human action recognition from skeletal data
• Efficient piecewise training of deep structured models for semantic segmentation
• Learning Multi-Domain Convolutional Neural Networks for Visual Tracking
• Large-scale Classification of Fine-Art Paintings: Learning The Right Metric on The Right Feature
• Guiding Long-Short Term Memory for Image Caption Generation
• 3D Morphable Models as Spatial Transformer Networks
• Evaluating color texture descriptors under large variations of controlled lighting conditions
• Bilinear CNNs for Fine-grained Visual Recognition
• Accurate Image Super-Resolution Using Very Deep Convolutional Networks
• How Important is Weight Symmetry in Backpropagation?
• RefineNet: Multi-Path Refinement Networks for High-Resolution Semantic Segmentation
• Super-Resolution via Image-Adapted Denoising CNNs: Incorporating External and Internal Learning
• Deeply Supervised Depth Map Super-Resolution as Novel View Synthesis
• A convolutional autoencoder approach for mining features in cellular electron cryo-tomograms and weakly supervised coarse segmentation
• Fast and Accurate Image Super-Resolution with Deep Laplacian Pyramid Networks
• REMAP: Multi-layer entropy-guided pooling of dense CNN features for image retrieval
• Deeply-Recursive Convolutional Network for Image Super-Resolution
• Deep Spatial Feature Reconstruction for Partial Person Re-identification: Alignment-Free Approach
• End-to-end Global to Local CNN Learning for Hand Pose Recovery in Depth Data
• Understanding trained CNNs by indexing neuron selectivity
• Boosting Occluded Image Classification via Subspace Decomposition Based Estimation of Deep Features
• Deep Spatial Pyramid: The Devil is Once Again in the Details
• Learning to Recognize 3D Human Action from A New Skeleton-based Representation Using Deep Convolutional Neural Networks
• When Face Recognition Meets with Deep Learning: an Evaluation of Convolutional Neural Networks for Face Recognition
• Sketch-a-Net that Beats Humans
• Deformable Object Tracking with Gated Fusion
• Deep Learning for the Classification of Lung Nodules
• Iris super-resolution using CNNs: is photo-realism important to iris recognition?
• A comparative study of 2D image segmentation algorithms for traumatic brain lesions using CT data from the ProTECTIII multicenter clinical trial
• Unsupervised High-level Feature Learning by Ensemble Projection for Semi-supervised Image Classification and Image Clustering
• Neonatal Pain Expression Recognition Using Transfer Learning
• Unsupervised ore/waste classification on open-cut mine faces using close-range hyperspectral data
• Learning a Discriminative Prior for Blind Image Deblurring
• Multi-Cue Zero-Shot Learning with Strong Supervision
• Deep Learning for Object Saliency Detection and Image Segmentation
• Evolutionary Synthesis of Deep Neural Networks via Synaptic Cluster-driven Genetic Encoding
• Learning Joint Feature Adaptation for Zero-Shot Recognition
• Learning and Recognizing Human Action from Skeleton Movement with Deep Residual Neural Networks
• Context-aware CNNs for person head detection
• Fast ConvNets Using Group-wise Brain Damage
• Optimization of Convolutional Neural Network using Microcanonical Annealing Algorithm
• Self-Supervised Learning for Spinal MRIs
• 3D Human Pose Estimation from a Single Image via Distance Matrix Regression
• On Face Segmentation, Face Swapping, and Face Perception
• Exploring Context with Deep Structured models for Semantic Segmentation
• Online Multiple Pedestrians Tracking using Deep Temporal Appearance Matching Association
• Automatic learning of gait signatures for people identification
• Im2Struct: Recovering 3D Shape Structure from a Single RGB Image
• Deep Semantic Face Deblurring
• Discriminative Label Consistent Domain Adaptation
• Simple and Efficient Learning using Privileged Information
• ASP Vision: Optically Computing the First Layer of Convolutional Neural Networks using Angle Sensitive Pixels
• LightNet: A Versatile, Standalone Matlab-based Environment for Deep Learning
• Deep Motion Features for Visual Tracking
• Synthesizing Dynamic Patterns by Spatial-Temporal Generative ConvNet
• Deep SR-ITM: Joint Learning of Super-Resolution and Inverse Tone-Mapping for 4K UHD HDR Applications
• One-to-many face recognition with bilinear CNNs
• Deep Laplacian Pyramid Networks for Fast and Accurate Super-Resolution
• Selective Deep Convolutional Features for Image Retrieval
• Self-Supervised Video Representation Learning With Odd-One-Out Networks
• On Classification of Distorted Images with Deep Convolutional Neural Networks
• From Selective Deep Convolutional Features to Compact Binary Representations for Image Retrieval
• Dense Recurrent Neural Networks for Scene Labeling
• Bottom-Up and Top-Down Reasoning with Hierarchical Rectified Gaussians
• Learning Fully Convolutional Networks for Iterative Non-blind Deconvolution
• Large Margin Structured Convolution Operator for Thermal Infrared Object Tracking
• Accelerating Deep Learning with Shrinkage and Recall
• Supervised Adversarial Networks for Image Saliency Detection
• On the Importance of Normalisation Layers in Deep Learning with Piecewise Linear Activation Units
• Beyond Deep Residual Learning for Image Restoration: Persistent Homology-Guided Manifold Simplification
• One Network to Solve All ROIs: Deep Learning CT for Any ROI using Differentiated Backprojection
• Sparse Factorization Layers for Neural Networks with Limited Supervision
• Deep Shape Matching
• Feature Selection Convolutional Neural Networks for Visual Tracking
• Auto-ML Deep Learning for Rashi Scripts OCR
• Objects as context for detecting their semantic parts
• Peri-Net-Pro: The neural processes with quantified uncertainty for crack patterns
• Learning Image Matching by Simply Watching Video
• Robust Optimization for Deep Regression
• Joint Learning of Siamese CNNs and Temporally Constrained Metrics for Tracklet Association
• Scene Parsing via Dense Recurrent Neural Networks with Attentional Selection
• BPGrad: Towards Global Optimality in Deep Learning via Branch and Pruning
• Discovery Radiomics for Pathologically-Proven Computed Tomography Lung Cancer Prediction
• Deep neural network based sparse measurement matrix for image compressed sensing
• Seeing into Darkness: Scotopic Visual Recognition
• Deep learning based fence segmentation and removal from an image using a video sequence
• Latent Hinge-Minimax Risk Minimization for Inference from a Small Number of Training Samples