cse 803 fall 2008 stockman1 veggie vision by ibm ideas about a practical system to make more...

23
CSE 803 Fall 2008 Stockma n 1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery store.

Post on 19-Dec-2015

218 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 1

Veggie Vision by IBM

Ideas about a practical systemto make more efficient the selling

and inventory of produce in a grocery store.

Page 2: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 2

Problem is recognizing produce

Page 3: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 3

15+ years of R&D now

This information was shared by IBM researchers. Since that time, the system has been tested in small markets and has been modified according to that experience.

Page 4: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 4

Up to 400 produce types

Page 5: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 5

Page 6: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 6

Practical problems of application environment

Page 7: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 7

Engineering the solution

Page 8: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 8

System to operate inside the usual checkout station

• together with bar code scanner

• together with scale

• together with accounting

• together with inventory

• together with employee

• within typical store environment

* figure shows system asking for help from the cashier in making final decision on touch screen

Page 9: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 9

Modifying the scale

Page 10: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 10

Need careful lighting engineering

Page 11: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 11

Need to segment product from background, even through plastic

Page 12: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 12

Previously published thresholding decision

Page 13: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 13

Quality segmented image obtained

Page 14: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 14

Design of pattern recognition paradigm (from 1997)

FEATURES are: color, texture, shape, and size all represented uniformly by HISTOGRAMS

Histograms capture statistical properties of regions – any number of regions.

Page 15: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 15

Matching procedure Sample product represented by

concatenated histograms: about 400 D 350 produce items x 10 samples = 3500

feature vectors of 400D each Have about 2 seconds to compare an

unknown sample to 3500 stored samples (3500 dot products)

Analyze the k nearest: if closest 2 are from one class, recognize that class (sure)

Page 16: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 16

HSI for pixel color: 6 bits for hue, 5 for saturation and intensity

For each pixel

quantify H

HIST[H]++

same for S&I

Page 17: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 17

Histograms of 2 limes versus 3 lemons

Distribution or population concept adds robustness:

• to size of objects

• to number of objects

• to small variations of color (texture, shape, size)

Page 18: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 18

Texture: histogram results of LOG filter[s] on produce pixels

Leafy produce A

Leafy produce B

Page 19: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 19

Shape: histogram of curvature of boundary of produce

Page 20: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 20

Banana versus lemon or cucumber versus lime

Large range of curvatures indicates complex object

Small range of curvatures indicates roundish object

Page 21: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 21

Size is also represented by a histogram

Page 22: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 22

Each pixel gets a “size” as the minimum distance to boundary

Purple grapes Chinese eggplants

Page 23: CSE 803 Fall 2008 Stockman1 Veggie Vision by IBM Ideas about a practical system to make more efficient the selling and inventory of produce in a grocery

CSE 803 Fall 2008 Stockman 23

Learning and adaptation System “easy” to train: show it

produce samples and tell it the labels. During service: age out oldest

sample; replace last used sample with newly identified one.

When multiple labeled samples match the unknown, system asks cashier to select from the possible choices.