Detecting Potential Falling Objects by Inferring Human Action and - - PowerPoint PPT Presentation

▶
detecting potential falling objects
SMART_READER_LITE
LIVE PREVIEW

Detecting Potential Falling Objects by Inferring Human Action and - - PowerPoint PPT Presentation

Feb 27, 2023 36 likes •348 views

Detecting Potential Falling Objects by Inferring Human Action and Natural Disturbance Bo Zheng*, Yibiao Zhao* (*equal contribution) Joey C. Yu, Katsushi Ikeuchi , Song-Chun Zhu Goal- understand the potential falling objects Oh, i ts

slide-1
SLIDE 1

Detecting Potential Falling Objects by Inferring Human Action and Natural Disturbance

Bo Zheng*, Yibiao Zhao* (*equal contribution) Joey C. Yu, Katsushi Ikeuchi , Song-Chun Zhu

slide-2
SLIDE 2

Goal-understand the potential falling objects

“Oh, it’s dangerous!”

slide-3
SLIDE 3

Motivations

  • safety surveillance system,
  • children, elders and people with disabilities
  • Robotics -rescue

DARPA robotics

slide-4
SLIDE 4

Issues

  • Human can imagine but

machine cannot.

  • Doing the serious physical

simulation?

  • various collisions
  • large number of objects
  • huge variation in size,

shape, material

Time consuming!

slide-5
SLIDE 5

Related work

  • Physics reasoning – understand support relations
  • “Block world revisit” [Gupta, ECCV10]
  • Support relations inference [Silberman, ECCV12]
  • Blocks, Support, and Stability [Jia, CVPR13]
  • Support surface prediction [Hoiem, ICCV13]
  • Cognitive science – Interpret human’s thought
  • Probabilistic model [Hamrick, CogSc11]
  • Robotics – avoid the obstacles
  • Safe motion planning [Petti, IROS05]
slide-6
SLIDE 6

Outline

  • Introduction
  • motivation
  • related works
  • Method
  • Experimental results
  • Discussion
slide-7
SLIDE 7

Observation – causality of the falling risk

  • “Cause” – the physical disturbance (energy

absorbed)

  • “Result” – much uncontrolled energy released
slide-8
SLIDE 8

Physical reasoning Disturbance field

Pipeline

Physical reasoning 3D scene Stability Motion disturbance Statistics Disturbance observation

Risk evalu ation

slide-9
SLIDE 9

Risk evaluation

slide-10
SLIDE 10

Part I: Physical reasoning

slide-11
SLIDE 11

Physical reasoning

[CVPR13, Zheng]

slide-12
SLIDE 12

Step 1: Segmentation

[CVPR13, Zheng]

slide-13
SLIDE 13

Step 2: volumetric completion

Segmentation result Result of volumetric completion

Holes Holes filled [CVPR13, Zheng]

slide-14
SLIDE 14

Step 3: Stability optimization

[SWC sampling, A. Babu’ 03] [CVPR13, Zheng]

slide-15
SLIDE 15

Part II: Risk evaluation

slide-16
SLIDE 16

Physical risk definition

Potential energy

disturbance

Absorb energy Release energy

slide-17
SLIDE 17

Example

slide-18
SLIDE 18

Disturbance field by human activities

Primary motion Secondary motion

slide-19
SLIDE 19

Disturbance field by human activities

slide-20
SLIDE 20

Other disturbances

human activity wind earthquake

slide-21
SLIDE 21

Outline

  • Introduction
  • motivation
  • related works
  • overview
  • Method
  • Experimental results
  • Discussion
slide-22
SLIDE 22

Point cloud segmentation Our method

Results of physical reasoning

slide-23
SLIDE 23

Results of physical reasoning

Point cloud segmentation Our method

slide-24
SLIDE 24

Large scale indoor scene

slide-25
SLIDE 25

Risk evaluation

slide-26
SLIDE 26

Under different disturbances

slide-27
SLIDE 27

input

  • utput
slide-28
SLIDE 28
slide-29
SLIDE 29

Outline

  • Introduction
  • motivation
  • related works
  • overview
  • Method
  • Experimental results
  • Discussion
slide-30
SLIDE 30

Discussion: Human v.s. Machine?

  • There is no ground truth
  • People have big variance on safety understanding
slide-31
SLIDE 31

Thank you for your attention!