I've forked this in order to adapt darknet with a very minor tweak that outputs the coordinates of the objects detected. I'm using the advice from Brian O'Donnell on stack exchange.
This repository contains a suite of R packages which perform image algorithms currently not available in other R packages like magick, imager or EBImage.
These algorithms are put into different packages because of license differences. Currently the following R packages are available:
| Package | Functionality | License | Details |
|---|---|---|---|
| image.CornerDetectionF9 | FAST-9 corner detection for images | BSD-2 | Details |
| image.CornerDetectionHarris | Harris corner detection for images | BSD-2 | Details |
| image.LineSegmentDetector | Line Segment Detector (LSD) for images | AGPL-3 | Details |
| image.ContourDetector | Unsupervised Smooth Contour Line Detection for images | AGPL-3 | Details |
| image.CannyEdges | Canny Edge Detector for Images | GPL-3 | Details |
| image.Otsu | Otsu's Image Segmentation Method | MIT | Details |
| image.dlib | Speeded up robust features (SURF) and histogram of oriented gradients (HOG) features | AGPL-3 | Details |
| image.libfacedetection | CNN for Face Detection | BSD-3 | Details |
| image.darknet | Image classification using darknet with deep learning models AlexNet, Darknet, VGG-16, Extraction (GoogleNet) and Darknet19. As well object detection using the state-of-the art YOLO detection system | MIT | Details |
| image.OpenPano | Image Stitching | see file LICENSE | Details |
| image.DenoiseNLMeans | Non-local means denoising | see file LICENSE | Details |
More packages and extensions are under development.
A presentation given at the useR-2017 conference is available in file presentation-user2017.pdf
- Some packages are on CRAN
install.packages("image.CannyEdges")
install.packages("image.ContourDetector")
install.packages("image.CornerDetectionF9")
install.packages("image.CornerDetectionHarris")
install.packages("image.dlib")
install.packages("image.libfacedetection")
install.packages("image.LineSegmentDetector")
install.packages("image.Otsu")
install.packages("image.binarization")
- You can see if the binary packages for your operating system are on the BNOSAC drat repo at https://github.com/bnosac/drat
- If they are, you can just install them as follows, where you replace
thepackagenamewith one of the packages you are interested in
install.packages("thepackagename", repos = "https://bnosac.github.io/drat")
Install the development version of packages as follows:
install.packages("remotes")
remotes::install_github("bnosac/image", subdir = "image.CornerDetectionF9")
remotes::install_github("bnosac/image", subdir = "image.CornerDetectionHarris")
remotes::install_github("bnosac/image", subdir = "image.LineSegmentDetector")
remotes::install_github("bnosac/image", subdir = "image.ContourDetector")
remotes::install_github("bnosac/image", subdir = "image.CannyEdges")
remotes::install_github("bnosac/image", subdir = "image.Otsu")
remotes::install_github("bnosac/image", subdir = "image.dlib")
remotes::install_github("bnosac/image", subdir = "image.darknet")
remotes::install_github("bnosac/image", subdir = "image.DenoiseNLMeans")
remotes::install_github("bnosac/image", subdir = "image.libfacedetection")
remotes::install_github("bnosac/image", subdir = "image.OpenPano")
Need support in image recognition? Contact BNOSAC: http://www.bnosac.be