Saliency and KAZE Features Authors: Siddharth Srivastava, Prerana - - PowerPoint PPT Presentation

▶
saliency and kaze features
SMART_READER_LITE
LIVE PREVIEW

Saliency and KAZE Features Authors: Siddharth Srivastava, Prerana - - PowerPoint PPT Presentation

Nov 14, 2022 16 likes •329 views

Adaptive Image Compression Using Saliency and KAZE Features Authors: Siddharth Srivastava, Prerana Mukherjee, Dr. Brejesh Lall Department of Electrical Engineering Indian Institute of Technology, Delhi SPCOM 2016 Overview Introduction

slide-1
SLIDE 1

Adaptive Image Compression Using Saliency and KAZE Features

Authors: Siddharth Srivastava, Prerana Mukherjee,

  • Dr. Brejesh Lall

SPCOM 2016

Department of Electrical Engineering Indian Institute of Technology, Delhi

slide-2
SLIDE 2

Overview

  • Introduction
  • Background
  • Proposed Methodology
  • Experimental Results and Discussions
  • Conclusion

SPCOM 2016

slide-3
SLIDE 3

Introduction

  • Primary Objectives

– To achieve image compression based

  • n

importance of the contents in the image. – Utilize properties from local regions to identify region importance.

SPCOM 2016

slide-4
SLIDE 4

Introduction

  • Underlying Principles

– Differed Importance: The human eye gives more importance to salient regions – Transition: The surrounding regions around salient

  • bjects make them distinguishable

– Rejection: Perceptually less significant Pixels/Regions can be made further insignificant – Object Characterization

  • Well defined boundary
  • Distinctive appearance
  • Uniqueness

SPCOM 2016

slide-5
SLIDE 5

Introduction

  • Core Methodology

– We adapt the quality factor in JPEG compression scheme for each block instead of a global quality factor – This adaptation of quality parameter is based on saliency map and strength of KAZE keypoints

SPCOM 2016

slide-6
SLIDE 6

Introduction

  • Key Contributions:

– JPEG Compatible – First to introduce KAZE for image compression – Maintains better perceptual quality at high compression ratios

SPCOM 2016

slide-7
SLIDE 7

Saliency Map Generation

Extraction of features from the Image Form Activation Maps based on those features Combining Maps for different features into one

Background: Saliency Map

SPCOM 2016

slide-8
SLIDE 8

Background: Saliency Map Computation

Aimed at segmenting objects Weighted Combination

SPCOM 2016

  • P. Mukherjee, B. Lall, and A. Shah, “Saliency map based improved segmentation,” in Image Processing (ICIP), 2015 IEEE International

Conference on. IEEE, 2015.

slide-9
SLIDE 9

Background: KAZE

  • KAZE is a recent feature detection technique

which exploits the non linear scale space to detect keypoints along edges and sharp discontinuities.

  • Non linear diffusion filtering allows KAZE to

achieve less blurring on edges as compared to Gaussian Blurring.

SPCOM 2016

slide-10
SLIDE 10

Saliency Map and KAZE

  • Fig. 1: (a) Original Image (b) Saliency Map (c) KAZE keypoints

SPCOM 2016

a) b) c)

slide-11
SLIDE 11

Proposed Methodology

  • Fig. 2: Architecture of (a) Compression and (b) Decompression

SPCOM 2016

slide-12
SLIDE 12

Adapting Quality Parameter (QP) with Saliency Response

i: the 8x8 block in the image

  • M. T. Khanna, K. Rai, S. Chaudhury, and B. Lall, “Perceptual depth preserving saliency based image compression,” in Proceedings
  • f the 2nd International Conference on Perception and Machine Intelligence. ACM, 2015, pp. 218–223.

SPCOM 2016

slide-13
SLIDE 13

Adapting Quality Parameter (QP) with Keypoint Response

i: the 8x8 block in the image

SPCOM 2016

slide-14
SLIDE 14

Algorithm I: Algorithm for Piecewise Adaptive QP

– QSal = QJPEG * salBoost – MR = mean(image_response) – BS = blockStrength

β1 * QSal if BS = 0 and salBoost < 1 QSal if BS = 0 and salBoost ≥ 1

  • r BS ≥ α * MR and BS < (1-α) *

MR (1-α) * QSal if α * MR (1+α) * QSal if BS < (1-α)*MR and BS ≤ (1+α)*MR β2 * QSal

  • therwise

boostedQP =

*0 < α ≤ 0.5, β1 < 1 and β2 > 1

SPCOM 2016

slide-15
SLIDE 15

Results

SPCOM 2016

  • Fig. 3: (a) Plot showing the change in PSNR rate as the compression ratio

changes (b) Plot between FSIMc with the varying compression ratio

slide-16
SLIDE 16

Results

SPCOM 2016

  • Fig. 4: (a) Original Image (b) Results after JPEG compression

(c) Results after Adaptive Compression (Proposed Approach) a) b) c)

slide-17
SLIDE 17

References

  • P. Mukherjee, B. Lall, and A. Shah, “Saliency map based

improved segmentation,” in Image Processing (ICIP), 2015 IEEE International Conference on. IEEE, 2015.

  • M. T. Khanna, K. Rai, S. Chaudhury, and B. Lall,

“Perceptual depth preserving saliency based image compression,” in Proceedings of the 2nd International Conference on Perception and Machine Intelligence. ACM, 2015, pp. 218–223.

  • F. Alcantarilla, A. Bartoli, and A. J. Davison, “Kaze

features,” in Computer Vision–ECCV 2012. Springer, 2012, pp. 214–227.

SPCOM 2016

slide-18
SLIDE 18

SPCOM 2016

slide-19
SLIDE 19

SPCOM 2016

slide-20
SLIDE 20

KAZE: Background

SPCOM 2016

slide-21
SLIDE 21

KAZE: Background

SPCOM 2016

slide-22
SLIDE 22

KAZE: Background

SPCOM 2016

slide-23
SLIDE 23

KAZE: Background

SPCOM 2016

slide-24
SLIDE 24

KAZE: Background

SPCOM 2016

slide-25
SLIDE 25

equation for building non linear scale space using AOS

KAZE: Keypoint Detection

SPCOM 2016

slide-26
SLIDE 26

Comparison between gaussian blurring and nonlinear diffusion

Non linear vs linear scale space

SPCOM 2016

slide-27
SLIDE 27

Feature detection

KAZE: Keypoint Detection

SPCOM 2016

slide-28
SLIDE 28

Scharr edge filter The Scharr operator is the most common technique with two kernels used to estimate the two dimensional second derivatives horizontally and vertically. The operator for the two direction is given by the following formula:

KAZE: Keypoint Detection

SPCOM 2016

slide-29
SLIDE 29

Feature description

KAZE: Keypoint Description

SPCOM 2016

slide-30
SLIDE 30

KAZE: Keypoint Description

SPCOM 2016

slide-31
SLIDE 31

Piecewise function for further adapting QP

SPCOM 2016