Removed double spaces after period in documents for consistency

build.gradle

@@ -35,7 +35,7 @@ project.ext.jsr250_version = '1.0'
 project.ext.jsr305_version = '3.0.2'
 project.ext.trove4j_version = '3.0.3'
 
-// Which native platforms are supported can be specified in the command line.  Otherwise
+// Which native platforms are supported can be specified in the command line. Otherwise
 // the default is to support all of them
 if (project.hasProperty("native_arch")) {
     project.ext.native_arch = [project.native_arch]

integration/boofcv-all/readme.txt

@@ -1,2 +1 @@
-The purpose of this package is to create an easy way to include ALL of BoofCV.  This includes packages in
-integration
+The purpose of this package is to create an easy way to include ALL of BoofCV. This includes packages in integration

integration/boofcv-android/examples/video/readme.md

@@ -3,7 +3,7 @@ capturing images and displaying the results. Much of the tedious handling of the
 life cycle, image formats, and other book keeping has been taken care of for a more streamlined 
 development of computer vision applications.
 
-Support for Camera 1 API has been deprecated and will eventually be removed from boofCV.  
+Support for Camera 1 API has been deprecated and will eventually be removed from boofCV.
 For Camera 1 API it's recommended that you use BoofCV v0.29 or earlier.
 
 # Performance

main/boofcv-core/readme.txt

@@ -1,2 +1,2 @@
-The purpose of this package is to create an easy way to include all of BoofCV.  It is dependent on all the packages
+The purpose of this package is to create an easy way to include all of BoofCV. It is dependent on all the packages
 in main.

main/boofcv-feature/feature_todo.txt

@@ -20,7 +20,7 @@
  * Add sub-pixel using a polynomial fit
 
 - Canny Edge Detector
- * Uses generalized algorithm for non-max right now.  It should use a specialized one for speed that traces
+ * Uses generalized algorithm for non-max right now. It should use a specialized one for speed that traces
    the edge intensity and direction
 
 - Add whole image detection capability to SSD detectors
@@ -35,5 +35,5 @@
   * Used round to reduce round off error?  Right now it is equivalent of floor
 
 
-- Add an escape value to score associate.  if the value computed so far isn't better than X, stop.
+- Add an escape value to score associate. if the value computed so far isn't better than X, stop.
   Suggested by David G Ortega

main/boofcv-ip/TODO.TXT

@@ -18,7 +18,7 @@
    - Provide support for GrayS32?
    - Do a better job normalizing?
    - See ShowImageDerivative
- * Check comments for Sobel and Prewitt.  Can you really compute it from those kernels?
+ * Check comments for Sobel and Prewitt. Can you really compute it from those kernels?
  * Include the image border for the Laplacian
 
 - Denoise
@@ -51,11 +51,11 @@
  * Sparse Kernel Convolve for when there is a lot of zeros in the kernel
    - Unroll
    - Compare speed to hand derived kernels in gradient
- * Special unroll for symmetric kernels.  need to store fewer elements.
+ * Special unroll for symmetric kernels. need to store fewer elements.
 
 - Kernels
  * Should Gaussian kernels be provided with only the exp() part?
-   - For numerical stability.  These are not app
+   - For numerical stability. These are not app
  * Support kernels with an even size
  * Support different width and height
 

main/boofcv-ip/design_issues.txt

@@ -4,9 +4,9 @@ Contains a discussion of design issues that there is no clear solution for and r
 --------------- Image Derivatives ----------------------------------------
 
  * For image derivatives with a built in blur (sobel, prewitt, gaussian) the blur portion should
-   be normalized to 1.  If it is not then the image intensity is being scaled up, exaggerating the magnitude.
+   be normalized to 1. If it is not then the image intensity is being scaled up, exaggerating the magnitude.
  * Disadvantage for normalizing in integer images is that you are effective reducing the precision of the
-   output due to discretization.  Also the division will slow the filter down.
+   output due to discretization. Also the division will slow the filter down.
  * For integer this is not done, and might not be done for all floating point kernels.
  * It is standard practice not to do this for integer images.
  * Will also exaggerate the affects of numerical overflow for integer and could affect feature detectors.
@@ -26,8 +26,8 @@ Possible Solution:
 
 Non-Maximum Suppression
 
-Regions can have pixels with identical values that are peaks.  Typically in synthetic regions
-cause by symmetry.  No detections will happen there.  Can use relaxed non-max to detect features in
-those regions, but the "same" feature is detected multiple times.  Could fix the problem by creating
-a list, sort by intensity and removing all but one duplicate.  Not going to implement since I don't
+Regions can have pixels with identical values that are peaks. Typically in synthetic regions
+cause by symmetry. No detections will happen there. Can use relaxed non-max to detect features in
+those regions, but the "same" feature is detected multiple times. Could fix the problem by creating
+a list, sort by intensity and removing all but one duplicate. Not going to implement since I don't
 think this is a significant issue.

main/checks/README.TXT

@@ -1,4 +1,4 @@
 The point of the checks directory is so that modules which are independent of each other
-can have their interactions tested in one location.  This was inspired by needing to
-check to see if data structures in feature can be correctly serialized by IO.  However
+can have their interactions tested in one location. This was inspired by needing to
+check to see if data structures in feature can be correctly serialized by IO. However
 IO and feature are completely independent of each other.

misc/planning.txt

@@ -250,23 +250,23 @@ Future
 - Fiducial Square
   * Regression tests that can determine if orientation is consistent
     - If multiple squares on same surface see if their orientation is the same
-  * Image fiducial.  Look at all the images and compute pixel weighting of pattern based on variance.
+  * Image fiducial. Look at all the images and compute pixel weighting of pattern based on variance.
                      Put all patterns which meet the correctness fraction then use the weighted sum to
                      select the winner.
   * For is black border black test, ignore the outer and inner most pixels to reduce false negative rate
-  * Consider alternative algorithm for small (or all) squares.  Contour -> gradient -> cluster/hough
+  * Consider alternative algorithm for small (or all) squares. Contour -> gradient -> cluster/hough
   * When optimizing the edges don't sample points tangent to the line sample points parallel to the adjacent sides
-    that way it will sample inside the square more often.  Maybe interpolate between slope of both sides?
-  * Look at alternative methods for generatng interpolated inner image.  bilinear is probably exagerating.  surface integral?
+    that way it will sample inside the square more often. Maybe interpolate between slope of both sides?
+  * Look at alternative methods for generatng interpolated inner image. bilinear is probably exagerating. surface integral?
 
-- Look at OpenCV's watershed example.  It's actually a useful technique!
+- Look at OpenCV's watershed example. It's actually a useful technique!
 
 - Background Model
   * Shadow removal
   * remove support for Planar once Interleaved is supported enough
 
 - ImplSsdCorner_S16
-  * Handle the border.  Currently is is just skipped
+  * Handle the border. Currently is is just skipped
   * If the image is resized that will cause problems since the border can affect the location of local
     maximums
   * Create a generic function for handling borders?  Give it a function and it returns a solution?

settings.gradle

@@ -26,5 +26,5 @@ include 'examples','demonstrations','applications',
 if (System.getenv()['ANDROID_HOME']) {
     include 'integration:boofcv-android'
 } else {
-    logger.warn('Skipping integration/android because ANDROID_HOME has not been set!  See integration/android/readme.txt')
+    logger.warn('Skipping integration/android because ANDROID_HOME has not been set! See integration/android/readme.txt')
 }