2016年斯普林格计算机视觉评价方法全书

需积分: 10 67 浏览量更新于2024-07-19 收藏 26.18MB PDF 举报

《2016年Springer计算机视觉度量》是一本由Scott Krig编著的专业教材，专为科学与工程学科提供全面的学术课程框架。该书以"Computer Vision Metrics"为名，旨在系统地调查、分类并分析计算机视觉领域、视觉神经系统以及深度学习的理论和技术。它不仅包含了实践操作和指导性的内容，还对这些领域的核心概念进行了深入剖析。本书共分为两个版本，纸质版的国际标准书号(ISBN)为978-3-319-33761-6，电子版则为978-3-319-33762-3，并获得了DOI 10.1007/978-3-319-33762-3的认可。在版权方面，图书符合Springer International Publishing的出版标准，所有内容受版权保护，包括但不限于翻译、重印、图像再利用、朗诵、广播、微缩复制、物理方式的复制品制作，以及任何形式的信息存储、检索、电子改编、计算机软件开发等。书中涵盖了广泛的通用术语、注册商标和服务标志，这些都是知识产权的重要组成部分。通过阅读这本书，读者可以了解到计算机视觉技术的基础理论、关键性能指标、评估方法，以及这些技术如何与视觉科学的最新研究成果相结合，以及深度学习模型在处理视觉任务中的应用和优化策略。对于计算机视觉研究人员、工程师、学生以及对该领域感兴趣的专业人士来说，这是一本不可多得的参考资料和学习工具，有助于提升他们在这个快速发展的领域的理解和实践能力。

Synthetic Interest Point Alphabet . . . .................. 266

Hybrid Synthetic Overlays on Real Images . . . . . . . . . . . . . 268

Summary . . ...................................... 269

Chapter 7: Le arning Assignments . . . ................... 271

8 Vision Pipelines and Optimizations ................... 273

Stages, Operations, and Resources . . . . . . . . . . . . . . . . . . . . . 274

Compute Resource Budgets . . ........................ 275

Compute Units, ALUs, and Accelerators . . ............. 277

Power Use ..................................... 278

Memory Use ................................... 278

I/O Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282

The Vision Pipeline Examples . . . . . . . . . . . . . . . . . . . . . . . . 282

Automobile Recognition . . ......................... 282

Face, Emotion, and Age Recognition . . . . . . . . .......... 289

Image Classiﬁcation . ............................. 296

Augmented Reality . .............................. 299

Acceleration Alternatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304

Memory Optimizations . . .......................... 304

Coarse-Grain Parallelism . . ........................ 307

Fine-Grain Data Parallelism ........................ 308

Advanced Instruction Sets and Accelerators . . . . . . . . . . . . . 310

Vision Algorithm Optimizations and Tuning . . . . . . . . . . . . . . 311

Compiler and Manual Optimizations . . . . . . . . . . . . . . . . . . 312

Tuning . . . ..................................... 312

Feature Descriptor Retroﬁt, Detectors,

Distance Functions ............................... 313

Boxlets and Convolution Acceleration . . . . . . . . ......... 313

Data-Type Optimizations, Integer vs. Float ............. 314

Optimization Resources ............................. 314

Summary . . ...................................... 315

Chapter 8: Le arning Assignments . . . ................... 316

9 Feature Learning Architecture Taxonomy

and Neuroscience Background ....................... 319

Neuroscience Inspirations for Computer Vision . . . . . . . . . . . . 320

Feature Generation vs. Feature Learning ................. 321

Terminology of Neuroscience Appli ed to Computer Vision . . . 322

Classes of Fe ature Learning . . . . . . . . . . . . . . . . . . . . . . . . . . 327

Convolutional Feature Weight Learning ................ 328

Local Feature Descriptor Learning ................... 328

Basis Feature Composition and Dictionary Learning ....... 329

Summary Perspective on Feature Learning Methods ....... 329

Machine Learning Models for Computer Vision . . . ......... 329

Expert Systems .................................. 330

Statistical and Mathematical Ana lysis Methods .......... 331

Neural Science Inspired Methods . . . . . . . . . . . . . . . . . . . . 331

Deep Learning .................................. 331

DNN Hacking and Misclassiﬁcation ................ 333

Contents xv

History of Machine Learning (ML) and Feature Learning ..... 333

Historical Survey, 1940s–2010s ...................... 334

Artiﬁcial Neural Network (ANN) Taxonomy Overview .... 338

Feature Learning Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 339

Learned Feature Descriptor Types . . . . . . . . . . . . . . . . . . . . 339

Hierarchical Feature Learning ....................... 340

How Many Feat ures to Learn? ...................... 340

The Power Of DNNs . . . .......................... 341

Encoding Efﬁciency . ............................. 341

Handcrafted Features vs. Handcrafted Deep Learning . ..... 341

Invariance and Robustness Attributes for Feature

Learning ...................................... 343

What Are the Best Features and Learning

Architectures? .................................. 343

Merger of Big Data, Analytics, and Computer Vision . . . . . . 345

Key Technology Enablers . . ........................ 347

Neuroscience Concepts .............................. 348

Biology and Blueprint . ............................ 349

The Elusive Uniﬁed Learning Theory . . . . . . . . . . . . . . . . . 350

Human Visual System Architecture . . . . . . . . . . . . . . . . . . . 351

Taxonomy of Feature Learning Architectures ............. 356

Architecture Topolo gies ........................... 357

ANNs (Artiﬁcial Neural Networks) . . . . . . . . . . . . . . . . . 358

FNN (Feed Forward Neural Network) . . . . . .......... 358

RNN (Recurrent Neural Network) . . ................ 358

BFN (Basis Function Network) . . . ................. 359

Ensembles, Hybrids . . . . ........................ 359

Architecture Compon ents and Layers . . . . . . . . . . . . . . . . . 360

Layer Totals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360

Layer Connection Topology ...................... 361

Memory Model ............................... 362

Training Protocols . . ........................... 362

Input Sampling Methods . . ....................... 363

Dropout, Reconﬁguration, Regularization . . . . . . . . . . . . 363

Preprocessing, Numeric Conditioning . . . . . . . . . . . . . . . 365

Features Set Dimensions . . ....................... 365

Feature Initialization . . ......................... 366

Features, Filters . . ............................. 366

Activation, Transfer functions . . . . . . . . . . . . . . . . . . . . . 367

Post-processing, Numeric Conditioning . . . . . . . . . . . . . . 368

Pooling, Subsampling, Downsampling, Upsampling . . . . . 369

Classiﬁers . . . . ............................... 371

Summary . . ...................................... 371

Chapter 9: Le arning Assignments . . . ................... 373

xvi Contents

10 Feature Learning and Deep Learning Architecture

Survey ......................................... 375

Architecture Survey . ............................... 376

FNN Architecture Survey . ......................... 377

P—Perceptron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377

MLP, Multilayer Perceptron, Cognitron, Neocognitron . . . 383

Concepts for CNNs, Convnets, Deep MLPs . .......... 387

LeNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417

AlexNet, ZFNet ............................... 419

VGGNet and Variants MSRA-22, Baidu Deep Image,

Deep Residual Learning . . . . . . . . . . . . . . . . . . . . . . . . . 422

Half-CNN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425

NiN, Maxout . . . . . . . . ......................... 426

GoogLeNet, InceptionNet . . . . . . . . . . . . . . . . . . . . . . . . 431

MSRA-22, SPP-Net, R-CNN, MSSNN, Fast-R-CNN . . . . 434

Baidu, Deep Im age, MINWA . . ................... 437

SYMNETS—Deep Symmetry Networks . . . . . . . . . . . . . 438

RNN Architecture Survey .......................... 442

Concepts for Recurrent Neural Networks . . . . . . . . . . . . . 443

LSTM, GRU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451

NTM, RNN-NTM, RL-NTM . . . .................. 454

Multidimensional RNNs, MDRNN ................. 457

C-RNN, QDRNN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460

RCL-RCNN .................................. 461

dasNET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463

NAP—Neural Abstraction Pyramid . . . . ............. 465

BFN Architecture Survey . ......................... 469

Concepts for Machine Learning and Basis Feature

Networks . . . . . ............................... 469

PNN—Polynomial Neural Network, GMDH . ......... 486

HKD—Kernel Descriptor Learning . . . . ............. 488

HMP—Sparse Feat ure Learning ................... 490

HMAX and Neurological Models .................. 495

HMO—Hierarchical Model Optimization . . . . . . . . . . . . 506

Ensemble Methods . . . . . . . . . . . . ..................... 506

Deep Neural Network Futures . . . ...................... 508

Increasing Depth to the Max—Deep Residual

Learning (DRL) . . . .............................. 509

Approximating Complex Models Using A Simpler

MLP (Model Compression) . . . . . . . . . . . . . . . . . . . . . . . . . 510

Classiﬁer Decomposition and Recombination . . . . . . . . . . . . 511

Summary . . ...................................... 511

Chapter 10: Learning Assignments . . ................... 513

Appendix A: Synthetic Feature Analysis ................... 515

Appendix B: Survey of Ground Truth Datasets .............. 547

Appendix C: Imag ing and Computer Vision Resources ........ 555

Contents xvii

Image Capture and Representation

“The changing of bodies into light, and light into bodies, is very conformable to the course

of Nature, which seems delighted with transmutations.”

—Isaac Newton

Computer vision starts with images. This chapter surveys a range of topics dealing with capturing,

processing, and representing images, including computational imaging, 2D imaging, and 3D depth

imaging methods, sensor processing, depth-ﬁeld processing for stereo and monocular multi-view

stereo, and surface reconstruction. A high-level overview of selected topics is provided, with

references for the interested reader to dig deeper. Readers with a strong background in the area of

2D and 3D imaging may beneﬁt from a light reading of this chapter.

Image Sensor Technology

This section provides a basic overview of image sensor technology as a basis for understanding how

images are formed and for developing effective strategies for image sensor processing to optimize the

image quality for computer vision.

Typical image senso rs are created from either CCD cells (charge-coupled device) or standard

CMOS cells (complementary metal-oxide semiconductor). The CCD and CMOS sensors share

similar characteristics and both are widely used in commercial cameras. The majority of sensors

today use CMOS cells, though, mostly due to manufacturing considerations. Sensors and optics are

often integrated to create wafer-scale cameras for applications like biology or microscopy, as shown

in Fig. 1.1.

Image sensors are designed to reach speciﬁc design goals with different applications in mind,

providing varying levels of sensit ivity and quality. Consult the manufacturer’s information to get

familiar with each sensor. For example, the size and material composition of each photodiode sensor

cell element is optimized for a given semiconductor manufacturing process so as to achieve the best

trade-off between silicon die area and dynamic response for light intensity and colo r detection.

For computer vision, the effects of sampling theory are relevant—for example, the Nyquist

frequency applied to pixel coverage of the target scene. The sensor resolution and optics together

must provide adequate resolution for each pixel to image the features of interest, so it follows that a

feature of interest should be imaged or sampled at least two times greater than the minimum size of

the smallest pixels of importance to the feature. Of course, 2 oversampling is just a minimum target

for accuracy; in practice, single pixel wide features are not easily resolved.

Springer International Publishing Switzerland 2016

S. Krig, Computer Vision Metrics, DOI 10.1007/978-3-319-33762-3_1

剩余652页未读，继续阅读

OMG59E9

粉丝: 1
资源: 4

2016年斯普林格计算机视觉评价方法全书

Computer Vision Metrics, Survey, Taxonomy and Analysis | Scott Krigs

Computer Vision Metrics 无水印pdf

Computer Vision Metrics.pdf

Real-Time Vision for Human-Computer Interaction

Springer - Geometric Algebra For Computer Graphics - 2008

Springer - Geometric Algebra for Computer Graphics (2008)

Springer-fibers-and-Delta-Conj

Springer-Handbook-of-Speech-Processing

Springer-Verlag Curves and Surfaces for Computer Graphics.pdf

Springer---SystemVerilog-for-Design--2nd-Edition._VHDL/FPGA/Verilog_Others_

最新资源