Download presentation
Presentation is loading. Please wait.
Published byDorthy Cross Modified over 9 years ago
1
Business Identification: Spatial Detection Alexander Darino Week 5
2
Outline Recap of Previous Work Business Name Detection Business Name Matching Business Spatial Detection Weaknesses to Current Approach Alternatives to Current Approach Acknowledgements 2
3
Outline Latitude Longitude Geocoding Reverse Geocoding Nearby Businesses ImageOCR Detected Text Business Name Matching Business Identification Business Spatial Detection Week 4Week 5 3
4
Previous Work 4 ImageWhere Am I? Latitude, Longitude Geocoding Reverse Geocoding Nearby Businesses 65 George S Aiken Co Winghart's Burger & Whiskey Bar Market Square Bella Sera On the Square Chipotle NOLA Las Velas …
5
Business Name Detection Latitude Longitude Geocoding Reverse Geocoding Nearby Businesses Image OCR Detected Text Business Name Matching Business Identification Business Spatial Detection 5
6
Business Name Detection 6
7
… … 7
8
Business Name Matching Latitude Longitude Geocoding Reverse Geocoding Nearby Businesses ImageOCR Detected Text Business Name Matching Business Identification Business Spatial Detection 8
9
Business Name Matching Developed Confidence Attribution Algorithm – Confidence of OCR Token being Name Token Example: Confidence of “ESTUANT” representing “RESTAURANT” Point-based system – Confidence of Name appearing in Image Sum of points of matching OCR Text Use logarithmically-normalized points to determine business inclusion threshold 9
10
Business Name Matching 10
11
11
12
Business Name Matching 12
13
13
14
Business Name Matching 14
15
Business Name Matching 15
16
Business Name Matching 16
17
Business Name Matching 17 Note: This originally did not appear because it did not exceed the confidence threshold. It now appears because it contributes to the Business Name Identification
18
Business Spatial Identification Latitude Longitude Geocoding Reverse Geocoding Nearby Businesses ImageOCR Detected Text Business Name Matching Business Identification Business Spatial Detection Business Spatial Detection 18
19
Business Spatial Identification 19
20
Business Spatial Identification 20 Aiken George S Co Category:Food, Grocery Address:218 Forbes Ave Pittsburgh, PA 15222 Phone: (412) 391-6358 Rating: 4.5/5 (2 Reviews) Category:Food, Grocery Address:218 Forbes Ave Pittsburgh, PA 15222 Phone: (412) 391-6358 Rating: 4.5/5 (2 Reviews)
21
Business Spatial Identification 21
22
Business Spatial Identification 22
23
Business Spatial Identification 23 Bruegger's Bagels Category:Bagels Address:Market Sq Pittsburgh, PA 15222 Phone: (412) 281-2515 Rating: Not Rated Category:Bagels Address:Market Sq Pittsburgh, PA 15222 Phone: (412) 281-2515 Rating: Not Rated
24
Weaknesses to Current Approach Latitude Longitude Geocoding Reverse Geocoding Nearby Businesses Image OCR Detected Text Business Name Matching Business Identification Business Spatial Detection 24
25
Weaknesses to Current Approach Lots of Garbage 25
26
Weaknesses to Current Approach Fragmented Word Detection 26
27
Weaknesses to Current Approach Fails with non-orthogonal perspective 27 Did I already mention lots of garbage?
28
Weaknesses to Current Approach Fails with non-roman text Not scale-invariant 28
29
ALTERNATIVE APPROACHES Two different 29
30
Alternative #1: Image Matching Latitude Longitude Geocoding Reverse Geocoding Nearby Businesses Image Match to Storefront Image Business Identification Business Spatial Detection 30
31
Alternative #1: Image Matching 31
32
Alternative #1: Image Matching Weaknesses – Storefront images aren’t always available for matching – Computationally Expensive Hundreds of images to compare to – Nothing new – Boring! 32
33
Alternative #2: Template Matching Latitude Longitude Geocoding Reverse Geocoding Nearby Businesses Image Render Templates of Business Names in Different Fonts Business Spatial Detection Image Matching (eg. SIFT, HAAR) Image Matching (eg. SIFT, HAAR) Template Images Business Identification 33
34
Alternative #2: Template Matching Tambellini 34
35
Alternative #2: Template Matching OCR Not Scale Invariant Unbounded Search Fragmented Recognition Roman-only font Alternative #2 Scale Invariant Bounded Search Whole-word recognition All fonts 35
36
Acknowledgements Subh – Provided several ideas regarding template matching using SIFT, HAAR features, etc 36
37
Thank You
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.