You are on page 1of 15

SELECTIVELY FILTERING IMAGE FEATURES USING A PERCENTAGE OCCUPANCY HIT-OR-MISS TRANSFORM

Paul Murray and Stephen Marshall

Overview
The Hit-or-Miss Transform (HMT) The Percentage Occupancy Hit-or-Miss Transform (POHMT) The Percentage Occupancy Plot (PO Plot) Properties of the PO Plot A Method for Selectively Filtering Image Features Examples Conclusions

The Hit-or-Miss Transform


Morphological transform that can be designed for feature detection/pattern recognition
Foreground Probe

Background Probe

Uses two SEs which probe the image in search of a match One SE searches for places where it fits inside features in the foreground A second searches for places where it fits around features in the background A positive result is returned if and only if both SEs find a match at some point in the image

The Hit-or-Miss Transform

In some cases, noise prevents the probes from fitting a feature which should otherwise be matched
BFG

BBG

This can be overcome by allowing a percentage P < 100% of the probes to fit the image

The Percentage Occupancy Hit-or-Miss Transform


The POHMT may be computed by measuring the extent to which image features occupy the SEs A feature is detected by the POHMT if P% of the SEs are occupied by some feature
T=50
100 80 PO of SEs 60 40 20 POFG POBG PO of SEs Extent to which BFG and BBG are occupied 100 80 60 40 20 50 100 150 Intensity 200 250 0 0 50 100 150 Intensity 200 250 POFG POBG Extent to which BFG and BBG are occupied

T=150
0 0

P 100

P 96

The PO Plot

A PO plot is generated by plotting the percentage occupancy of both probes against each other The parameter P of the POHMT may be estimated by finding the value of the critical point in the plot

The PO Plot and the POHMT


A PO plot may be used to estimate P from training data before applying the POHMT
PO Plot 100 80 POBG 60 40 20
Top Left Cell Top Right Cell Bottom Left Cell Bottom Right Cell

P=70

0 0

20

40

60 POFG

80

100

One pair of SEs were designed to match all 4 cells in the image The PO plot was used to estimate the only parameter, P The POHMT accurately detects all 4 cells despite the noise in the image when P = 70%

Selective Filtering
P = 90% 2 cells P=75% 3 cells P = 96% 1 cell

PO Plot

100 80 POBG 60 40 20 0 0 20 40
Top Left Cell Top Right Cell Bottom Left Cell Bottom Right Cell

60 POFG

80

100

The PO plot can be used to estimate values of P such that some features are discarded by the POHMT Increasing the value of P makes the transform more strict leading to fewer cells being detected

Selective Filtering
The previous example shows a powerful but simple technique for selectively filtering image features It is obvious that increasing the strictness of the transform by increasing the value of P leads to fewer features being detected However, it is NOT easy to isolate any of the four cells in the image The following novel method allows any combination of the image features to be segmented using just one composite SE Lets try to isolate the cell in the bottom right of the image which occupies around 80% of the SEs

Selective Filtering

Locate Compute Overlay Set P high low any the a enough enough empty bounding bounding to bounding todetect box exclude boxes for the boxes computed each the feature feature connected these of forinterest of the are interest component previous the locations image in the of the image sought feature This leads to features other than the desired one to be detected Mark the position of any empty bounding boxes in the ouput image

Selective Filtering - Example


PO Plot 100 80 POBG

P = 95

P = 91

P = 88

60 40 20

P = 84

P = 80

P = 76

0 0

20

40

60 POFG

80

100

A complementary pair of SEs were designed to detect dice in an image A PO plot was generated for each face of the die using these SEs The critical points of each SE were computed to allow accurate setting of parameters for the selective filtering method using bounding boxes

Selective Filtering - Example

Isolating dice displaying 1

Isolating dice displaying 6

Isolating dice displaying 3

Isolating dice displaying 2 or 5

Selective Filtering - Example


PO Plot 100 80 POBG 60 40 20 0 0 20 40 60 POFG 80 100

Noise was added to the image The PO plot was regenerated The selective filtering algorithm was used to locate dice displaying a value of 2

Summary
The HMT is a powerful method which can be used for for template matching The HMT is unreliable for processing noisy data The POHMT is an extension of the HMT which offers improved performance for processing noisy data The PO plot is a novel design tool which can be used to estimate parameters for the POHMT and similar transforms The PO plot can be exploited to make the POHMT capable of selectively filtering image features A novel approach which uses bounding boxes to locate features of interest has been demonstrated This approach allows features to be isolated in an image based on the extent to which they occupy a complementary pair of SEs

You might also like