pliman 2.0.0

New functions

• analyze_objects_iter() to execute an interactive section of analyze_objects().

• measure_disease_byl() to measure disease severity 'by leaf' in an image with several leaves.

• object_split() to split multiples objects of an image into a list of images.

• pca(), plot.pca(), get_biplot() as helper functions to perform Principal Component Analysis.

• rownames_to_column(), column_to_rownames(), separate_col(), round_cols() as helper functions to manipulate data.

• A set of poly_*() function to analyze polygons. All of them are based on a set of coordinate points describing the edge of the object(s). See ?utils_polygon for more details.

• get_wd_here() and set_wd_here() to deal with working directories.

• apply_fun_to_imgs() to apply a function (or functions) to a set of images stored in the working directory.

• make_brush(), make_mask(), and image_segment_mask() to create masks and segment images based on such a mask.

• image_segment_manual(), image segment kmeans(), and image_segment_mask() to perform image segmentation in different ways.

• A new family of efourier_*() functions to performs Elliptical Fourier Analysis.

  • efourier(): Elliptical Fourier Analysis
  • efourier_coefs(): Get Fourier coefficients
  • efourier_error(): Erros between the original and reconstructed outline
  • efourier_inv(): Inverse Elliptical Fourier Analysis
  • efourier_norm(): Normalized Fourier coefficients
  • efourier_power(): Power in Fourier Analysis
  • efourier_shape(): Draw shapes based on Fourier coefficients

• A new family of landmarks_*() functions to handle landmarks

  • landmarks(): Create image landmarks
  • landmarks_add(): Artificially inflates the number of landmarks
  • landmarks_angle(): Angles between landmarks
  • landmarks_dist(): Distances between landmarks
  • landmarks_regradi(): Pseudolandmarks with equally spaced angles

• object_edge() to detect edges in images using Sobel-Feldman Operator.

• A new family of *_shp() functions to analyze shape files.

  • image_shp() to construct a shape file from an image.
  • object_split_shp() to splits image objects based on a shapefile.
  • analyze_objects_shp() to analyze objects using shapefiles.
  • measure_disease_shp() to measure disease using shapefiles.

• New plot_index() function to plot an image index using raster package, and optionaly using the mapview package to show the image index.

• New image_view() function to create an interactive map view of an image. This function allows users to interactively edit and analyze an image using mapview and mapedit packages.

• New image_prepare_mv() function to prepare an image to be analyzed for analyze_objects_shp(). This function aligns and crops the image using either base or mapview visualization.

New features

• New viewer option added. Now, iterative functions such as pick_palette() and measure_disease_iter() have an argument viewer. If not provided, the value is retrieved using get_pliman_viewer(). This option controls the type of viewer to use for interactive plotting. The available options are "base" and "mapview". If set to "base", the base R graphics system is used for interactive plotting. If set to "mapview", the mapview package is used, allowing the users to draw shapes like points and polygons with mapedit package. To set this argument globally for all functions in the package, you can use the set_pliman_viewer() function. For example, you can run set_pliman_viewer("mapview") to set the viewer option to "mapview" for all functions.

• Haralick's features that quantify pixel texture for image objects were included.

• Several measures were added in analyze_objects(). The function now wraps some poly_*() functions to compute shape measures such as width, length, elongation, circularity. Haralick's features are now computed by default. . This improvement was at cost of a slight increase in computation time.

• analyze_objects(), measure_disease(), and measure_disease_byl() have now a filter argument that applies a median filtering in the binary mask. This is useful to reduce the noise in the segmentation of objects.

• Arguments reference_larger and reference_smaller were included in analyze_objects() indicating when the larger/smaller object in the image must be used as the reference object.

• Arguments efourier and nharm included in analyze_objects(). If efourier = TRUE, Elliptical Fourier analysis is computed for each object depending on the number of harmonics (nharm).

• Logical arguments reference_larger and reference_smaller included in analyze_objects(). Those indicates when the larger/smaller object in the image must be used as the reference object. This only is valid when reference = TRUE and reference_area indicates the area of the reference object. IMPORTANT. When reference_smaller is used, objects with an area smaller than 1% of the mean of all the objects are ignored. This is used to remove possible noise in the image such as dust. So, be sure the reference object has an area that will be not removed by that cutpoint.

• Rcpp and RcppArmadillo dependencies were included, allowing the implementation of C++ code. This will dramatically reduce the time computing of some functions/procesures. As an example, we wave.

  • Reduction in time processing from more than 5 minutes to less than 1 second using the new object_rgb() function to extract the RGB values from an image (1445 x 1084) with ~1400 objects.
  • Reduction in time processing of the set of *_poly() functions.

Minor changes

• get_measures() now remove known objects from the results when using the id argument.
• The right-hand of argument measure in get_measures() now accepts a numeric object stored in the global environment.
• analyze_objects() now returns the objects object_index and object_rgb when the argument pattern is used. Thanks to João Paulo Oliveira Ribeiro for alerting me regarding this issue.
• New argument reference in analyze_objects() to adjust measures using a reference object in the image.
• Argument object_index in analyze_objects() now recognizes the names of built-in indexes (see ?pliman_indexes()).
• plot.image_index() not limits the number of pixels to reduce plotting time.
• show_image argument changed with plot to standardize the argument across functions.
• rgb_to_hsb() optimized using C++.
• Change rows and cols with nrow and ncol, respectively, in functions analyze_objects_shp(), image_shp(), measure_disease_shp(), and object_split_shp(), to standardize the arguments across functions.