cropFins: cropFins

In haimeh/finFindR: Fin Identification

Usage

cropFins(imageName, cropNet, workingImage, saveDir, minXY = 100, target = 1,
  threshold = 0.45, dim = c(225, 150, 3, 1))

Arguments

imageName	path to image to be cropped
cropNet	mxnet model for isolating dolphin
workingImage	workspace for holding image during processing
saveDir	directory to save cropped image
minXY	minimum width/height in pixels for valid crop
target	label indicating layer representing class
threshold	threshold value for triggering a crop

Details

Function loads image and uses the given neural network to isolate any dorsal fins therin. The neural network returns a matrix whose values range [0,1]. Using a contiguous region algorithm (blob analysis), the x,y range for each excitation on the matrix is found. The range is then rescaled to find the analogous region in the original image. This region is cropped from said image and saved as a jpeg to the given directory.

haimeh/finFindR documentation built on July 17, 2021, 12:56 a.m.