Features from accelerated segment test: Difference between revisions

Revision as of 19:06, 4 April 2013

Revision as of 19:32, 4 April 2013

Line 55:

== Comparison with other detectors ==

In Rosten's research,<ref>Edward Rosten, [http://lanl.arXiv.org/pdf/0810.2434 FASTER and better: A machine learning approach to corner detection]</ref> FAST and FAST-ER detector are evaluated on several different datasets alongand compared with other detectors. Those detectors includethe [[DoG]], [[Harris affine region detector|Harris]] detector, [[Harris-Laplace]], [[Shi-Tomasi]], and [[SUSAN]] andcorner parameter settings are as follows:detectors.

The parameter settings for the detectors (other than FAST) are as follows:

{| class="wikitable"

|-

! Detector !! Parameter Setting || Value

|-

| rowspan="4" | DoG

| DoG || Scales per octave 3

Initial blur σ 0.8

Octaves 4

|-

| DoG || Scales per octave 3

| SUSAN || Distance threshold 4.0

| 3

|-

| Initial blur σ 0.8

| 0.8

|-

| Octaves 4

| 4

|-

| SUSAN

| Harris, Shi-Tomasi || Blur σ 2.5

| SUSAN || Distance threshold 4.0

| 4.0

|-

| Harris, Shi-Laplace || Initial blur σTomasi 0.8

| Blur σ

Harris blur 3

| 2.5

|-

Octaves 4

| rowspan="4" | Harris-Laplace

| Initial blur σ

Scales per octave 10

| 0.8

|-

| Harris blur 3

| 3

|-

Octaves| 4Octaves

| 4

|-

| Scales per octave 10

| 10

|-

| General parameters |
| ε
| 5 pixels

|-

|}

Line 110 ⟶ 130:

! Detector !! Training set pixel rate !! Test set pixel rate

|-

| FAST n = 9 || 188 || 179

|-

| FAST n= 12 || 158 || 154

|-

| Original FAST n =12 || 79 || 82.2

|-

| FAST-ER || 75.4 || 67.5