calc.blob.stats: Calculate Color and Spatial Statistics from Blob Region
In imagefx: Extract Features from Images

Description Usage Arguments Details Value Author(s) See Also Examples

Wrapper function (see details) that calculates various statistics on x,y data that corresponds to an image.

1	calc.blob.stats(img, xy.coords)

`img`	Matrix whose values at `xy.coords` are statistically analyzed.
`xy.coords`	Index locations corresponding region of interest (e.g. blob region) in the `img`

Function calls multiple statistical functions (e.g. mean, sd) and applies them to regions in the img according to the index locations given by xy.coords. In general, this function is commented to promote any modifications needed to fit the users needs. For example, adding or removing statistical analyses is straight forward.

Numeric vector giving the statistics of the blob region.

Alex J.C. Witsil

mean sd sum colMeans rowMeans length skewness kurtosis

############
### EG 1 ###
############
## example with synthetic data

## create an image that is a simple gaussian
img <- build.gaus(100,100,sig.x=2,sig.y=10,x.mid=80,y.mid=60)

## find the where the maximum value is
img.max <- which(img==max(img),arr.ind=TRUE)

## define a sigma for the low pass filter
sig=5

## define the low pass filter as another gaussian
lp.filt <- build.gaus(nrow(img),ncol(img),sig.x=sig)

## define a window size for the connected component algorithm
win.size=0.05

## perform the blob detection
blob <- blob.extract(img=img, blob.point=img.max,win.size=win.size,gaus=lp.filt)

#################################
### CALCULATE BLOB STATISTICS ###
#################################

blob.stats <- calc.blob.stats(img, blob$xy.coords)
print(blob.stats)


############
### EG 2 ###
############

## example with volcano image data.
data(sakurajima)

######################
### PRE PROCESSING ###
######################

## crop accroding to these corner values
xleft = 1
xright = 188
ybottom = 1
ytop = 396

## crop the image using crop.image
cropped <- crop.image(sakurajima, xleft, ybottom, xright, ytop)

## redefine the crop image
img <- cropped$img.crop

## separate the image into red, green, and blue images
r.img <- img[,,1]
g.img <- img[,,2]
b.img <- img[,,3]

## remove the mean
r.img <- r.img-mean(r.img)
g.img <- g.img-mean(g.img)
b.img <- b.img-mean(b.img)

## calculate the the plane trend...
r.img.trend <- fit3d(r.img)
g.img.trend <- fit3d(g.img)
b.img.trend <- fit3d(b.img)

## remove the trend
r.img.dtrend <- r.img-r.img.trend
g.img.dtrend <- g.img-g.img.trend
b.img.dtrend <- b.img-b.img.trend


################################
### SET UP SOME FILTER MASKS ###
################################

## define a sigma for the LP Gaussian Filter
gaus.sig=30

## build the Gaussian filter
gaus <- build.gaus(nrow(img),ncol(img),gaus.sig)

## find the maximum value of each RGB channel
blob.r.point <- which(r.img.dtrend==max(r.img.dtrend),arr.ind=TRUE)
blob.g.point <- which(g.img.dtrend==max(g.img.dtrend),arr.ind=TRUE)
blob.b.point <- which(b.img.dtrend==max(b.img.dtrend),arr.ind=TRUE)

## set a window size to be used in the connected component algorithm
win.size = 0.05

## extract the blob xy locations
blob.r <- blob.extract(r.img.dtrend,blob.r.point,win.size,gaus)
blob.g <- blob.extract(g.img.dtrend,blob.r.point,win.size,gaus)
blob.b <- blob.extract(b.img.dtrend,blob.r.point,win.size,gaus)


#################################
### CALCULATE BLOB STATISTICS ###
#################################

r.blob.stats <- calc.blob.stats(r.img.dtrend, blob.r$xy.coords)
g.blob.stats <- calc.blob.stats(g.img.dtrend, blob.g$xy.coords)
b.blob.stats <- calc.blob.stats(b.img.dtrend, blob.b$xy.coords)

print(r.blob.stats)
print(g.blob.stats)
print(b.blob.stats)