kaze/orx
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add orx-jumpflood/README.md

Browse Source
This commit is contained in:
Edwin Jakobs
2019-11-29 12:43:24 +01:00
parent 50446eeb1b
commit 1c3411c587
2 changed files with 223 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@@ -12,7 +12,7 @@ A growing library of assorted data structures, algorithms and utilities.
- [`orx-filter-extension`](orx-filter-extension/README.md), `Program` extension method that provides Filter based `extend()` - [`orx-filter-extension`](orx-filter-extension/README.md), `Program` extension method that provides Filter based `extend()`
- [`orx-integral-image`](orx-integral-image/README.md), CPU-based and GPU-based implementation for integral images (summed area tables) - [`orx-integral-image`](orx-integral-image/README.md), CPU-based and GPU-based implementation for integral images (summed area tables)
- [`orx-interval-tree`](orx-interval-tree/README.md), data structure for accelerating point-in-interval queries. - [`orx-interval-tree`](orx-interval-tree/README.md), data structure for accelerating point-in-interval queries.
- `orx-jumpflood`, a filter/shader based implementation of the jump flood algorithm for finding fast approximate (directional) distance fields - [`orx-jumpflood`](orx-jumpflood/README.md), a filter/shader based implementation of the jump flood algorithm for finding fast approximate (directional) distance fields
- `orx-kdtree`, a kd-tree implementation for fast nearest point searches - `orx-kdtree`, a kd-tree implementation for fast nearest point searches
- [`orx-kinect-v1`](orx-kinect-v1/README.md), utilities to use Kinect V1 RGB-D sensors in OPENRNDR programs. - [`orx-kinect-v1`](orx-kinect-v1/README.md), utilities to use Kinect V1 RGB-D sensors in OPENRNDR programs.
- [`orx-mesh-generators`](orx-mesh-generators/README.md), triangular mesh generators - [`orx-mesh-generators`](orx-mesh-generators/README.md), triangular mesh generators

222
orx-jumpflood/README.md Normal file
View File

@@ -0,0 +1,222 @@
# orx-jumpflood
An OPENRNDR extra that provides GPU accelerated jump flooding functionality.
[Original jump flooding algorithm](https://www.comp.nus.edu.sg/~tants/jfa.html)
`orx-jumpflood` focusses on finding 2d distance and directional distance fields.
## Distance field example
`distanceFieldFromBitmap()` calculates distances to bitmap contours it stores
the distance in red and the original bitmap in green.
````kotlin
import org.openrndr.application
import org.openrndr.draw.*
import org.openrndr.extra.jumpfill.Threshold
import org.openrndr.extra.jumpfill.distanceFieldFromBitmap
import org.openrndr.ffmpeg.VideoPlayerFFMPEG
import org.openrndr.filter.blur.ApproximateGaussianBlur
fun main() = application {
configure {
width = 1280
height = 720
}
program {
val blurFilter = ApproximateGaussianBlur()
val blurred = colorBuffer(width, height)
val thresholdFilter = Threshold()
val thresholded = colorBuffer(width, height)
val distanceField = colorBuffer(width, height, type = ColorType.FLOAT32)
val videoCopy = renderTarget(width, height) {
colorBuffer()
}
val videoPlayer = VideoPlayerFFMPEG.fromDevice(imageWidth = width, imageHeight = height)
videoPlayer.play()
extend {
// -- copy videoplayer output
drawer.isolatedWithTarget(videoCopy) {
drawer.ortho(videoCopy)
videoPlayer.draw(drawer)
}
// -- blur the input a bit, this produces less noisy bitmap images
blurFilter.sigma = 9.0
blurFilter.window = 18
blurFilter.apply(videoCopy.colorBuffer(0), blurred)
// -- threshold the blurred image
thresholdFilter.threshold = 0.5
thresholdFilter.apply(blurred, thresholded)
distanceFieldFromBitmap(drawer, thresholded, result = distanceField)
drawer.isolated {
// -- use a shadestyle to visualize the distance field
drawer.shadeStyle = shadeStyle {
fragmentTransform = """
float d = x_fill.r;
if (x_fill.g > 0.5) {
x_fill.rgb = 1.0 * vec3(cos(d) * 0.5 + 0.5);
} else {
x_fill.rgb = 0.25 * vec3(1.0 - (cos(d) * 0.5 + 0.5));
}
"""
}
drawer.image(distanceField)
}
}
}
}
````
## Direction field example
`directionFieldFromBitmap()` calculates directions to bitmap contours it stores
x-direction in red, y-direction in green, and the original bitmap in blue.
```
import org.openrndr.application
import org.openrndr.draw.*
import org.openrndr.extra.jumpfill.Threshold
import org.openrndr.extra.jumpfill.directionFieldFromBitmap
import org.openrndr.ffmpeg.VideoPlayerFFMPEG
import org.openrndr.filter.blur.ApproximateGaussianBlur
fun main() = application {
configure {
width = 1280
height = 720
}
program {
val blurFilter = ApproximateGaussianBlur()
val blurred = colorBuffer(width, height)
val thresholdFilter = Threshold()
val thresholded = colorBuffer(width, height)
val directionField = colorBuffer(width, height, type = ColorType.FLOAT32)
val videoPlayer = VideoPlayerFFMPEG.fromDevice(imageWidth = width, imageHeight = height)
videoPlayer.play()
val videoCopy = renderTarget(width, height) {
colorBuffer()
}
extend {
// -- copy videoplayer output
drawer.isolatedWithTarget(videoCopy) {
drawer.ortho(videoCopy)
videoPlayer.draw(drawer)
}
// -- blur the input a bit, this produces less noisy bitmap images
blurFilter.sigma = 9.0
blurFilter.window = 18
blurFilter.apply(videoCopy.colorBuffer(0), blurred)
// -- threshold the blurred image
thresholdFilter.threshold = 0.5
thresholdFilter.apply(blurred, thresholded)
directionFieldFromBitmap(drawer, thresholded, result = directionField)
drawer.isolated {
// -- use a shadestyle to visualize the direction field
drawer.shadeStyle = shadeStyle {
fragmentTransform = """
float a = atan(x_fill.r, x_fill.g);
if (a < 0) {
a += 3.1415926535*2;
}
if (x_fill.g > 0.5) {
x_fill.rgb = 1.0*vec3(cos(a*1.0)*0.5+0.5);
} else {
x_fill.rgb = 0.25*vec3(cos(a*1.0)*0.5+0.5);
}
"""
}
drawer.image(directionField)
}
}
}
}
import org.openrndr.application
import org.openrndr.draw.*
import org.openrndr.extra.jumpfill.Threshold
import org.openrndr.extra.jumpfill.directionFieldFromBitmap
import org.openrndr.ffmpeg.VideoPlayerFFMPEG
import org.openrndr.filter.blur.ApproximateGaussianBlur
fun main() = application {
configure {
width = 1280
height = 720
}
program {
val blurFilter = ApproximateGaussianBlur()
val blurred = colorBuffer(width, height)
val thresholdFilter = Threshold()
val thresholded = colorBuffer(width, height)
val directionField = colorBuffer(width, height, type = ColorType.FLOAT32)
val videoPlayer = VideoPlayerFFMPEG.fromDevice(imageWidth = width, imageHeight = height)
videoPlayer.play()
val videoCopy = renderTarget(width, height) {
colorBuffer()
}
extend {
// -- copy videoplayer output
drawer.isolatedWithTarget(videoCopy) {
drawer.ortho(videoCopy)
videoPlayer.draw(drawer)
}
// -- blur the input a bit, this produces less noisy bitmap images
blurFilter.sigma = 9.0
blurFilter.window = 18
blurFilter.apply(videoCopy.colorBuffer(0), blurred)
// -- threshold the blurred image
thresholdFilter.threshold = 0.5
thresholdFilter.apply(blurred, thresholded)
directionFieldFromBitmap(drawer, thresholded, result = directionField)
drawer.isolated {
// -- use a shadestyle to visualize the direction field
drawer.shadeStyle = shadeStyle {
fragmentTransform = """
float a = atan(x_fill.r, x_fill.g);
if (a < 0) {
a += 3.1415926535*2;
}
if (x_fill.g > 0.5) {
x_fill.rgb = 1.0*vec3(cos(a*1.0)*0.5+0.5);
} else {
x_fill.rgb = 0.25*vec3(cos(a*1.0)*0.5+0.5);
}
"""
}
drawer.image(directionField)
}
}
}
}
```