Computer Vision and Image Analysis

7.

Deep Learning for Computer Vision

7.1.

Neural Network Foundations

7.1.1.

Artificial Neurons

7.1.1.1.

Perceptron Model

7.1.1.2.

Multi-layer Perceptrons

7.1.1.3.

Universal Approximation

7.1.2.

Activation Functions

7.1.2.1.

Sigmoid Function

7.1.2.2.

Hyperbolic Tangent

7.1.2.3.

ReLU and Variants

7.1.2.4.

Softmax Function

7.1.3.

7.1.3.1.

Mean Squared Error

7.1.3.2.

Cross-entropy Loss

7.1.3.3.

7.1.3.4.

Custom Loss Functions

7.1.4.

7.1.4.1.

Gradient Descent

7.1.4.2.

Stochastic Gradient Descent

7.1.4.3.

Mini-batch Training

7.1.4.4.

Learning Rate Scheduling

7.2.

Convolutional Neural Networks

7.2.1.

CNN Architecture

7.2.1.1.

Convolutional Layers

7.2.1.2.

7.2.1.3.

Fully Connected Layers

7.2.1.4.

Layer Connectivity

7.2.2.

Convolution Operation

7.2.2.1.

7.2.2.2.

Stride and Padding

7.2.2.3.

Parameter Sharing

7.2.2.4.

Translation Equivariance

7.2.3.

Pooling Operations

7.2.3.1.

7.2.3.2.

Average Pooling

7.2.3.3.

7.2.3.4.

Adaptive Pooling

7.2.4.

7.2.4.1.

Local Connectivity

7.2.4.2.

Hierarchical Features

7.2.4.3.

Receptive Fields

7.2.4.4.

7.3.

CNN Architectures

7.3.1.

Early Architectures

7.3.1.1.

7.3.1.2.

7.3.1.3.

Network in Network

7.3.2.

Deep Architectures

7.3.2.1.

7.3.2.2.

7.3.2.3.

Inception Modules

7.3.3.

Residual Networks

7.3.3.1.

Skip Connections

7.3.3.2.

ResNet Variants

7.3.3.3.

7.3.4.

Efficient Architectures

7.3.4.1.

7.3.4.2.

7.3.4.3.

7.4.

Training Deep Networks

7.4.1.

Data Preparation

7.4.1.1.

Data Augmentation

7.4.1.2.

7.4.1.3.

Dataset Splitting

7.4.2.

7.4.2.1.

7.4.2.2.

Batch Normalization

7.4.2.3.

7.4.2.4.

7.4.3.

Advanced Optimizers

7.4.3.1.

7.4.3.2.

7.4.3.3.

7.4.3.4.

Learning Rate Adaptation

7.4.4.

Transfer Learning

7.4.4.1.

Pretrained Models

7.4.4.2.

Feature Extraction

7.4.4.3.

Fine-tuning Strategies

7.4.4.4.

Domain Adaptation

7.5.

Image Classification

7.5.1.

Single-label Classification

7.5.1.1.

7.5.1.2.

Output Layer Configuration

7.5.1.3.

Training Strategies

7.5.2.

Multi-label Classification

7.5.2.1.

Problem Formulation

7.5.2.2.

Loss Function Design

7.5.2.3.

Evaluation Metrics

7.5.3.

Few-shot Learning

7.5.3.1.

7.5.3.2.

Prototypical Networks

7.5.3.3.

Siamese Networks

7.6.

Object Detection

7.6.1.

Two-stage Detectors

7.6.1.1.

7.6.1.2.

Region Proposal Networks

7.6.1.3.

Feature Pyramid Networks

7.6.2.

Single-stage Detectors

7.6.2.1.

YOLO Architecture

7.6.2.2.

7.6.2.3.

7.6.3.

Detection Components

7.6.3.1.

Anchor Generation

7.6.3.2.

Bounding Box Regression

7.6.3.3.

Classification Heads

7.6.3.4.

Non-maximum Suppression

7.6.4.

Evaluation Metrics

7.6.4.1.

Average Precision

7.6.4.2.

Mean Average Precision

7.6.4.3.

Intersection over Union

7.7.

Semantic Segmentation

7.7.1.

Fully Convolutional Networks

7.7.1.1.

Architecture Design

7.7.1.2.

Upsampling Methods

7.7.1.3.

Skip Connections

7.7.2.

Encoder-Decoder Architectures

7.7.2.1.

7.7.2.2.

7.7.2.3.

7.7.3.

Advanced Techniques

7.7.3.1.

Atrous Convolution

7.7.3.2.

Pyramid Pooling

7.7.3.3.

Attention Mechanisms

7.7.4.

Instance Segmentation

7.7.4.1.

7.7.4.2.

7.7.4.3.

Panoptic Segmentation

Previous

6. Classical Object Recognition

Go to top

Next

8. Motion and Video Analysis