Computer Vision Basics

• Image Terminology

• Binary Operations

• Filtering

• Edge Operators

1
Digital Image Terminology:

0 0 0 0 1 0 0
pixel (with value 94)
0 0 1 1 1 0 0
0 1 95 96 94 93 92 its 3x3 neighborhood
0 0 92 93 93 92 92
0 0 93 93 94 92 93 region of medium
0 1 92 93 93 93 93 intensity
0 0 94 95 95 96 95
resolution (7x7)

• binary image
• gray-scale (or gray-tone) image
• color image
• multi-spectral image
• range image
• labeled image 2
The Three Stages of Computer Vision
• low-level

image image

• mid-level

image features

• high-level

features analysis

3
Low-Level

sharpening

blurring

4
Low-Level

Canny

original image edge image

Mid-Level

ORT

data
structure
circular arcs and line segments 5
edge image
Mid-level

K-means
clustering
(followed by
connected
component
analysis)

original color image regions of homogeneous color

data
structure
6
Low- to High-Level

low-level
edge image

mid-level

consistent
high-level line clusters

Building Recognition
7
Binary Image Analysis
• used in a variety of applications:
part inspection
riveting
fish counting
document processing

• consists of a set of image analysis operations

that are used to produce or process binary
images, usually images of 0’s and 1’s.

00010010001000
00011110001000
00010010001000
8
Example: red blood cell image
 Many blood cells are
separate objects
 Many touch – bad!
 Salt and pepper noise
from thresholding
 How useable is this
data?
 What operations are
needed to clean it up?

9
Useful Operations
1. Thresholding a gray-tone image

2. Determining good thresholds

3. Filtering with mathematical morphology

4. Connected components analysis

5. Numeric feature extraction

• location features
• gray-tone features
• shape features ...
10
Thresholding
 Background is black
 Healthy cherry is bright
 Bruise is medium dark
 Histogram shows two
cherry regions (black
background has been
removed)

pixel
counts

0 256
gray-tone values 11
Automatic Thresholding:
Otsu’s Method
Grp 1 Grp 2
Assumption: the histogram is bimodal

Method: find the threshold t that minimizes

the weighted sum of within-group variances
for the two groups that result from separating
the gray tones at value t.

Works well if the assumption holds.

12
Thresholding Example

original image pixels above threshold

13
 Mathematical Morphology
(Dilation, Erosion, Closing, Opening)

• Dilation

Dilation expands the connected sets of 1s of a binary image.

It can be used for

1. growing features

2. filling holes and gaps

14
• Erosion

Erosion shrinks the connected sets of 1s of a binary image.

It can be used for

1. shrinking features

2. Removing bridges, branches and small protrusions

15
Structuring Elements

A structuring element is a shape mask used in

the basic morphological operations.

They can be any shape and size that is

digitally representable, and each has an origin.

box
hexagon disk
something

box(length,width) disk(diameter)
16
 Dilation with Structuring Elements

The arguments to dilation and erosion are

1. a binary image B
2. a structuring element S

dilate(B,S) takes binary image B, places the origin

of structuring element S over each 1-pixel, and ORs
the structuring element S into the output image at
the corresponding position.

0000 dilate 0110

1
0110 0111
11
0000 0000
S
B origin BS
17
Erosion with Structuring Elements

erode(B,S) takes a binary image B, places the origin

of structuring element S over every pixel position, and
ORs a binary 1 into that position of the output image only if
every position of S (with a 1) covers a 1 in B.

origin
0 0 1 1 0 0 0 0 0 0
1 erode
0 0 1 1 0 0 0 1 1 0
1
0 0 1 1 0 0 0 1 1 0
1
1 1 1 1 1 0 0 0 0 0

B S B S
18
 Opening and Closing

• Closing is the compound operation of dilation followed

by erosion (with the same structuring element)

• Opening is the compound operation of erosion followed

by dilation (with the same structuring element)

19
20
Application:
Gear Tooth Inspection
original
binary
image

detected
defects

21
Connected Components Labeling
Once you have a binary image, you can identify and
then analyze each connected set of pixels.

The connected components operation takes in a binary image

and produces a labeled image in which each pixel has the
integer label of either the background (0) or a component.

original thresholded opening+closing components

22
 Methods for CC Analysis
1. Recursive Tracking (almost never used)

2. Parallel Growing (needs parallel hardware)

3. Row-by-Row (most common)

a. propagate labels down to the bottom,

recording equivalences

b. Compute equivalence classes

c. Replace each labeled pixel with the

label of its equivalence class.
23
Labelings shown as Pseudo-Color

connected
components
of 1’s from
cleaned,
thresholded
image

connected
components
of cluster
labels

24