Using Deep Learning and Google Street View to Estimate the - - PowerPoint PPT Presentation

▶
using deep learning and google street view to estimate
SMART_READER_LITE
LIVE PREVIEW

Using Deep Learning and Google Street View to Estimate the - - PowerPoint PPT Presentation

Feb 03, 2024 268 likes •450 views

Using Deep Learning and Google Street View to Estimate the Demographic Makeup of the US Timnit Gebru, Jonathan Krause, Yilun Wang, Duyun Chen, Jia Deng, Erez Lieberman Aiden, Li Fei-Fei Presented by Chia-Wen Cheng Wed Nov 8, 2017 Each year,

slide-1
SLIDE 1

Using Deep Learning and Google Street View to Estimate the Demographic Makeup of the US

Timnit Gebru, Jonathan Krause, Yilun Wang, Duyun Chen, Jia Deng, Erez Lieberman Aiden, Li Fei-Fei Presented by Chia-Wen Cheng Wed Nov 8, 2017

slide-2
SLIDE 2

Each year, the U.S. Census Bureau spends $1 billion surveying the population.

slide-3
SLIDE 3

Challenges of Population Survey

  • Labor-intensive
  • Time-consuming
  • Ignore smaller areas
slide-4
SLIDE 4

A faster, more efficient, and higher-resolution way of studying the population?

slide-5
SLIDE 5

The type of car people own is a strong indicator of their demographic information.

slide-6
SLIDE 6

Vehicular Census via Google Street View Images

  • 200 American cities
  • 50 million Street View Images
  • 22 million vehicles
  • 2,657 different categories of cars
  • Vehicle characteristics

○ Make, model, year, body type...

slide-7
SLIDE 7

Automated System for Monitoring Demographic Trends

Street View images Car Detection Car Classification Extract car-related attributes Demographic Estimation

race, income, education, voting pattern

slide-8
SLIDE 8

Car Detection

  • Deformable Part Models (DPMs)
  • Tradeoff between performance and efficiency

Image credit: P. Felzenszwalb et al.

slide-9
SLIDE 9

Car Classification

Street View images Product shot images AlexNet 2657-way classification domain adaptation

slide-10
SLIDE 10

Car-Related Attributes

88 attributes:

  • The average number of detected cars per image
  • Average car price
  • Miles per gallon
  • Percent of total cars that are hybrids
  • Percent of total cars that are electric
  • Percent of total cars that are from each of seven countries
  • Percent of total cars that are foreign (not from the USA)
  • Percent of total cars from each of 11 body types
  • Percent of total cars whose year fall within each of five year ranges: 1990-1994, 1995-1999,

2000-2004, 2005-2009, and 2010-2014

  • Percent of total cars whose make is each of 58 makes in our dataset
slide-11
SLIDE 11

Demographic Estimation

88 car-related attributes

Ridge regression Logistic regression

Income Voter preference Race Education

slide-12
SLIDE 12

Results

Race

slide-13
SLIDE 13

Results

Education Income

slide-14
SLIDE 14

An interesting finding

“ If the number of sedans encountered during a 15-minute drive through a city is higher than the number of pickup trucks, the city is likely to vote for a Democrat during the next Presidential election; otherwise, it is likely to vote Republican. “

slide-15
SLIDE 15

Strengths

Overall:

  • Solve a practical problem
  • Think outside of the box

Technique:

  • Almost real-time monitoring
  • Fine spatial resolution
slide-16
SLIDE 16

Weaknesses

  • Hand-crafted car-related attributes
  • Correlation between car ownership and demographics ?
  • Generalizable to other demographic factors?

e.g. religion, birth rate, death rate, marriage age, marital status

slide-17
SLIDE 17

Extension

  • Other types of imagery

e.g. Drone View images, satellite images

  • Predict traffic flow