Review SketchNet: Sketch Classification with Web Images [CVPR `16] - - PowerPoint PPT Presentation

▶
review
SMART_READER_LITE
LIVE PREVIEW

Review SketchNet: Sketch Classification with Web Images [CVPR `16] - - PowerPoint PPT Presentation

Dec 02, 2022 785 likes •966 views

Review SketchNet: Sketch Classification with Web Images [CVPR `16] (Speaker. Doheon Lee) Problem in previous sketch-based image retrieval People have different sketch style Large difference btw sketch and image Manual Annotation

slide-1
SLIDE 1

1

Review

  • SketchNet: Sketch Classification with Web Images

[CVPR `16] (Speaker. Doheon Lee)

  • Problem in previous sketch-based image retrieval
  • People have different sketch style
  • Large difference btw sketch and image
  • Manual Annotation is expensive
  • Solution
  • Weakly supervised Learning
  • Triplet pair (anchor sketch, positive & negative images)
  • Sketch Net: S-Net (sketch), R-Net (image), and C-Net
  • C-Net: merge feature maps btw image and sketch
slide-2
SLIDE 2

Age Progression/Regression by Conditional Adversarial Autoencoder [CVPR `17]

20189008 Ben Jung (정병의)

slide-3
SLIDE 3

3

Table of Contents

  • Introduction
  • Problems of Previous Works
  • Main Idea & Solution: CAAE
  • Experiment & Result
  • Overview
slide-4
SLIDE 4

4

Introduction

  • Age Progression & Regression

Given face

35 years old 10 20 40 50

Progression/Aging Regression

slide-5
SLIDE 5

5

Problems of Previous Works

  • Group-wised learning
  • Query with label
  • Step-by-step transition

5 years old 10 30 60 !" !# !$

… …

query with label 10

slide-6
SLIDE 6

6

Main Idea

  • Group-wised learning è Joint learning
  • Query with label è Query without label
  • Step-by-step transition

è One-step & bidirectional transition query

slide-7
SLIDE 7

7

Main Idea: Manifold Traversing

  • Assumptions
  • The faces lies on a manifold (!)
  • Clustered by ages and personality
  • Traversing on the manifold corresponds to age/personality

transformation

slide-8
SLIDE 8

8

Solution: CAAE

  • Conditional Adversarial Autoencoder
slide-9
SLIDE 9

9

Solution: CAAE

  • Conditional Adversarial Autoencoder

!"

Uniform noise Real faces

!#$%

slide-10
SLIDE 10

10

Solution: CAAE

  • Effect of Discriminator on z (!")
slide-11
SLIDE 11

11

Solution: CAAE

  • Effect of Discriminator on image (!"#$)

without !"#$ with !"#$ without !"#$ with !"#$ without !"#$ with !"#$

slide-12
SLIDE 12

12

Experiment & Result

slide-13
SLIDE 13

13

Experiment & Result

  • Comparison with prior work
slide-14
SLIDE 14

14

Experiment & Result

  • Comparison with ground truth
slide-15
SLIDE 15

15

Overview

  • CAAE (Conditional Adversarial Autoencoder)
  • Manifold Traversing
  • Joint learning
  • Query without label
  • One-step & bidirectional transition
  • Discriminator on z
  • Discriminator on image
slide-16
SLIDE 16

16

THANK YOU