Skip to content

D
dlib
Commit 7f61b773 authored by Davis King's avatar Davis King
Browse files
Options

updated docs

parent e90fc107
Hide whitespace changes
Inline Side-by-side
Showing with 381 additions and 2 deletions
docs/docs/faq.xml
View file @ 7f61b773
......@@ -5,6 +5,41 @@
<title>Frequently Asked Questions</title>
<!-- ************************************************************************* -->
<questions group="Python">
<question text="Why is dlib.image_window missing?">
If you are getting the error <tt>module 'dlib' has no attribute 'image_window'</tt>
it is because you compiled dlib without GUI support (or you are using a copy of
dlib someone else compiled and they built it without GUI support). So note that
it is possible to compile dlib with any GUI tools. Some people want to do this because
they run dlib on systems that don't have any kind of GUI framework installed.
<p>
But since you are reading this you obviously want to use the GUI tools. The solution
is to get a copy of dlib and <tt>python setup.py install</tt> it yourself. It's easy.
Note that you might get a warning in the build output about X11 not being installed. Maybe that's
why you are getting this error in the first place. <b>READ THAT MESSAGE AND FOLLOW ITS INSTRUCTIONS</b>
since it tells you what to do to fix this.
</p>
</question>
<question text="Why is some function missing from the dlib module?">
If you are missing dlib.image_window then read <a href="#Whyisdlib.imagewindowmissing">the FAQ about that</a>.
If you are missing any other function then <b>it's because you are using an old version of dlib
that just doesn't have that function</b>. You need to install a newer version
of dlib. Please don't post questions about this on any of dlib's forums or
email me about it. Just install a new dlib. The only way to use features in
a new version of dlib is to get the new version of dlib. Often people think they have the
new version of dlib installed when really they have some old version installed.
You can see what version of dlib you are using by checking <tt>dlib.__version__</tt>.
</question>
</questions>
<questions group="General">
<question text="Why do I get USER_ERROR__inconsistent_build_configuration__see_dlib_faq_1?">
......
docs/docs/imaging.xml
View file @ 7f61b773
......@@ -101,6 +101,7 @@
<item>scan_image</item>
<item>scan_image_movable_parts</item>
<item>find_points_above_thresh</item>
<item>find_peaks</item>
<item>scan_image_pyramid</item>
<item>scan_image_boxes</item>
<item>scan_fhog_pyramid</item>
......@@ -144,6 +145,9 @@
</sub>
</item>
<item>find_bright_keypoints</item>
<item>find_dark_keypoints</item>
<item>hog_image</item>
<item>extract_fhog_features</item>
<item>fine_hog_image</item>
......@@ -166,10 +170,14 @@
<item>hysteresis_threshold</item>
<item>sobel_edge_detector</item>
<item>suppress_non_maximum_edges</item>
<item>normalize_image_gradients</item>
<item>remove_incoherent_edge_pixels</item>
<item>threshold_image</item>
<item>partition_pixels</item>
<item>hough_transform</item>
<item>image_gradients</item>
<item>find_bright_lines</item>
<item>find_dark_lines</item>
</section>
<section>
......@@ -186,6 +194,8 @@
<item>binary_difference</item>
<item>binary_complement</item>
<item>skeleton</item>
<item>encode_8_pixel_neighbors</item>
<item>find_line_endpoints</item>
</section>
<section>
......@@ -305,6 +315,34 @@
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>encode_8_pixel_neighbors</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/morphological_operations_abstract.h</spec_file>
<description>
This routine looks at the 8 pixels immediately surrounding a pixel
and encodes their on/off pattern into the 8 bits of an unsigned char.
Needless to say, this routine only makes sense for binary images.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>find_line_endpoints</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/morphological_operations_abstract.h</spec_file>
<description>
This routine finds endpoints of lines in a thinned binary image.
For example, if the image was produced by <a href="#skeleton">skeleton()</a>
or something like a Canny edge detector then you can use
find_line_endpoints() to find the pixels sitting on the ends of
lines.
</description>
</component>
<!-- ************************************************************************* -->
<component>
......@@ -504,6 +542,37 @@
<!-- ************************************************************************* -->
<component>
<name>find_bright_keypoints</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/edge_detector_abstract.h</spec_file>
<description>
This routine finds bright "keypoints" in an image. In general,
these are bright/white localized blobs. It does this by computing
the determinant of the image Hessian at each location and storing
this value into an output saliency image if both eigenvalues of
the Hessian are negative. If either eigenvalue is positive then
the saliency for that pixel is 0.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>find_dark_keypoints</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/edge_detector_abstract.h</spec_file>
<description>
This routine finds dark "keypoints" in an image. In general, these are dark
localized blobs. It does this by computing the determinant of the image
Hessian at each location and storing this value into an output saliency
image if both eigenvalues of the Hessian are positive. If either eigenvalue
is negative then the saliency for that pixel is 0.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>hog_image</name>
<file>dlib/image_keypoint.h</file>
......@@ -1404,6 +1473,36 @@
</examples>
</component>
<!-- ************************************************************************* -->
<component>
<name>find_bright_lines</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/edge_detector_abstract.h</spec_file>
<description>
This routine is similar to <a href="#sobel_edge_detector">sobel_edge_detector()</a>,
except instead of finding an edge it finds a bright/white line.
For example, the border between a black piece of paper and a white
table is an edge, but a curve drawn with a pencil on a piece of
paper makes a line.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>find_dark_lines</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/edge_detector_abstract.h</spec_file>
<description>
This routine is similar to <a href="#sobel_edge_detector">sobel_edge_detector()</a>,
except instead of finding an edge it finds a dark line.
For example, the border between a black piece of paper and a white
table is an edge, but a curve drawn with a pencil on a piece of
paper makes a line.
</description>
</component>
<!-- ************************************************************************* -->
<component>
......@@ -1466,6 +1565,35 @@
</component>
<!-- ************************************************************************* -->
<component>
<name>normalize_image_gradients</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/edge_detector_abstract.h</spec_file>
<description>
This function takes two gradient images produced by a function like
<a href="#sobel_edge_detector">sobel_edge_detector</a> and then
unit normalizes the gradient vectors.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>remove_incoherent_edge_pixels</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/edge_detector_abstract.h</spec_file>
<description>
This routine takes a list of edge pixels and discards the ones that
have outlying gradient directions. In particular, it finds the
largest group of pixels in an edge that are all within a user
specified angle of each other. Therefore, if you are looking for
straight lines then this routine is a useful tool since it allows
you to discard noisy edge pixels with inconsistent angles.
</description>
</component>
<!-- ************************************************************************* -->
<component>
......@@ -1944,6 +2072,21 @@
</component>
<!-- ************************************************************************* -->
<component>
<name>partition_pixels</name>
<file>dlib/image_transforms.h</file>
<spec_file link="true">dlib/image_transforms/thresholding_abstract.h</spec_file>
<description>
Finds a threshold value that would be reasonable to use with
<a href="#threshold_image">threshold_image</a>. It does this by
finding the threshold that partitions the pixels in an image into
two groups such that the sum of absolute deviations between each
pixel and the mean of its group is minimized.
</description>
</component>
<!-- ************************************************************************* -->
<component>
......@@ -2169,6 +2312,18 @@
</component>
<!-- ************************************************************************* -->
<component>
<name>find_peaks</name>
<file>dlib/image_processing.h</file>
<spec_file link="true">dlib/image_processing/scan_image_abstract.h</spec_file>
<description>
This routine finds all the points in an image with a pixel value
above a user supplied threshold that are also local maximizers.
</description>
</component>
<!-- ************************************************************************* -->
<component>
......
docs/docs/linear_algebra.xml
View file @ 7f61b773
......@@ -617,6 +617,7 @@
<item>drectangle</item>
<item>vector</item>
<item>point</item>
<item>dpoint</item>
<item>rotate_point</item>
<item>rotate_around_x</item>
<item>rotate_around_y</item>
......@@ -634,6 +635,10 @@
<item>rotation_matrix</item>
<item>get_rect</item>
<item>centered_rect</item>
<item>
<name>centered_rects</name>
<link>dlib/geometry/rectangle_abstract.h.html#centered_rects</link>
</item>
<item>set_aspect_ratio</item>
<item>set_rect_area</item>
<item>center</item>
......@@ -651,6 +656,13 @@
<item>distance_to_rect_edge</item>
<item>clip_line_to_rectangle</item>
<item>distance_to_line</item>
<item>line</item>
<item>signed_distance_to_line</item>
<item>reverse</item>
<item>intersect</item>
<item>count_points_on_side_of_line</item>
<item>count_points_between_lines</item>
</section>
......@@ -739,6 +751,81 @@
<!-- ************************************************************************* -->
<component>
<name>line</name>
<file>dlib/geometry.h</file>
<spec_file link="true">dlib/geometry/line_abstract.h</spec_file>
<description>
This object represents a line in the 2D plane. The line is defined by two
<a href="#dpoint">points</a> running through it. This object also includes a
unit normal vector that is perpendicular to the line.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>signed_distance_to_line</name>
<file>dlib/geometry.h</file>
<spec_file link="true">dlib/geometry/line_abstract.h</spec_file>
<description>
This function returns how far a point is from a <a href="#line">line</a>.
This is a signed distance. The sign indicates which side of the line the point
is on and the magnitude is the distance.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>reverse</name>
<file>dlib/geometry.h</file>
<spec_file link="true">dlib/geometry/line_abstract.h</spec_file>
<description>
This function returns a <a href="#line">line</a> object that represents the
same line but with the endpoints and normal vector flipped.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>intersect</name>
<file>dlib/geometry.h</file>
<spec_file link="true">dlib/geometry/line_abstract.h</spec_file>
<description>
This routine finds the point at the intersection of two <a href="#line">lines</a>.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>count_points_on_side_of_line</name>
<file>dlib/geometry.h</file>
<spec_file link="true">dlib/geometry/line_abstract.h</spec_file>
<description>
This routine takes a <a href="#line">line</a> and an array of points and counts how many
points are on one side of the line. Which side of the line is of interest is
selected by the user.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>count_points_between_lines</name>
<file>dlib/geometry.h</file>
<spec_file link="true">dlib/geometry/line_abstract.h</spec_file>
<description>
This routine takes a pair of <a href="#line">lines</a> and an array of points
and counts how many points are between the lines.
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>sparse_to_dense</name>
<file>dlib/sparse_vector.h</file>
......@@ -1370,6 +1457,19 @@
</description>
</component>
<!-- ************************************************************************* -->
<component>
<name>dpoint</name>
<file>dlib/geometry.h</file>
<spec_file link="true">dlib/geometry/vector_abstract.h</spec_file>
<description>
This object represents a point inside a Cartesian coordinate system.
Note that a dpoint is simply a typedef for a <a href="#vector">vector</a>
that is 2D and uses doubles to represent coordinate values.
</description>
</component>
<!-- ************************************************************************* -->
......
docs/docs/ml.xml
View file @ 7f61b773
......@@ -166,6 +166,10 @@ Davis E. King. <a href="http://jmlr.csail.mit.edu/papers/volume10/king09a/king09
<name>upsample</name>
<link>dlib/dnn/layers_abstract.h.html#upsample_</link>
</item>
<item>
<name>resize_to</name>
<link>dlib/dnn/layers_abstract.h.html#resize_to_</link>
</item>
<item>
<name>l2normalize</name>
<link>dlib/dnn/layers_abstract.h.html#l2normalize_</link>
......
docs/docs/release_notes.xml
View file @ 7f61b773
......@@ -11,6 +11,70 @@
<!-- ************************************************************************************** -->
<current>
New Features and Improvements:
- Deep Learning
- Added resize_to layer.
- Made loss_multiclass_log_per_pixel use CUDA. This greatly accelerates
training with this loss layer.
- Image Processing
- Added normalize_image_gradients() and remove_incoherent_edge_pixels().
- Added neighbors_24 (for use with label_connected_blobs())
- Added partition_pixels() and made threshold_image() use it to find the
default threshold if none is given. Also depreciated
auto_threshold_image() since using partition_pixels() to pick the
threshold is superior.
- Added overload of hysteresis_threshold() that uses partition_pixels()
to select thresholds.
- Added encode_8_pixel_neighbors() and find_line_endpoints().
- Added image_gradients, a tool for computing multi-scale first and
second order image gradients.
- Added find_bright_lines(), find_dark_lines(), find_bright_keypoints(),
and find_dark_keypoints().
- Added label_connected_blobs_watershed().
- Added find_peaks().
- Added find_pixels_voting_for_lines(), find_strong_hough_points(),
and perform_generic_hough_transform() to the hough_transform.
- Improved the stopping condition of the solve_qp_box_constrained() and
solve_qp_box_constrained_blockdiag() quadratic program solvers.
- Made find_min_single_variable() more numerically stable.
- Make visual studio build with all cores when building dlib.
- Added steal_memory() to the matrix.
- Improved the workflow for adding part annotations in imglab. You can now
use shift+click to rapidly add parts to an object. Each part will
automatically receive an integer label.
- Added .begin() and .end() to array2d.
- Added the line class and some related utility functions.
- Added centered_rects().
- Added numpy_image, which is a simple type safe interface to numpy arrays.
It is also a fully functional dlib image class, making interfacing dlib's
image processing to Python much cleaner.
Non-Backwards Compatible Changes:
- All CUDA code in dlib was moved to dlib/cuda.
Bug fixes:
- Replaced all the old Python C APIs interfacing to numpy with the new
numpy_image. This fixed some reference counting errors as a result.
- Fixed load_bmp() not loading certain types of BMP images.
- Fixed compile time bugs when trying to use CBLAS in some unusual environments.
- Fixed bug in global_function_search's constructor that takes initial function
evaluations. It wasn't assigning these values into the entire state of the
solver.
- Fixed compile time error in the matlab bindings. Also, fixed cell arrays
of complex types not binding correctly.
- Renamed BOOST_JOIN to DLIB_BOOST_JOIN to prevent name clashes when working
with boost.
- Fixed bug in hysteresis_threshold() that could cause incorrect output.
- The last release of dlib added more ODR violation checks. However, these
could erroneously trigger when using Visual Studio and CUDA in certain
workflows. This has been fixed.
- Fixed matrix objects not working correctly when sized greater than 4GB in
Visual Studio.
</current>
<!-- ************************************************************************************** -->
<old name="19.10" date="Mar 19, 2018">
New Features and Improvements:
- Deep Learning:
- Added scale_ layer, allowing implementation of squeeze-and-excitation networks.
......@@ -60,7 +124,7 @@ Bug fixes:
really bad training data or parameters have been added.
- Fixed the decayed running stats objects so they use unbiased estimators.
</current>
</old>
<!-- ************************************************************************************** -->
......
docs/docs/term_index.xml
View file @ 7f61b773
......@@ -149,6 +149,7 @@
<term file="dlib/dnn/layers_abstract.h.html" name="EXAMPLE_COMPUTATIONAL_LAYER_" include="dlib/dnn.h"/>
<term file="dlib/dnn/layers_abstract.h.html" name="fc_" include="dlib/dnn.h"/>
<term file="dlib/dnn/layers_abstract.h.html" name="con_" include="dlib/dnn.h"/>
<term file="dlib/dnn/layers_abstract.h.html" name="resize_to_" include="dlib/dnn.h"/>
<term file="dlib/dnn/layers_abstract.h.html" name="mult_prev_" include="dlib/dnn.h"/>
<term file="dlib/dnn/layers_abstract.h.html" name="extract_" include="dlib/dnn.h"/>
<term file="dlib/dnn/layers_abstract.h.html" name="upsample_" include="dlib/dnn.h"/>
......@@ -184,6 +185,7 @@
<term file="linear_algebra.html" name="resize_rect_width" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="resize_rect_height" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="centered_rect" include="dlib/geometry.h"/>
<term file="dlib/geometry/rectangle_abstract.h.html" name="centered_rects" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="set_aspect_ratio" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="set_rect_area" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="center" include="dlib/geometry.h"/>
......@@ -407,7 +409,7 @@
<term file="algorithms.html" name="square_root" include="dlib/algs.h"/>
<term file="linear_algebra.html" name="vector" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="point" include="dlib/geometry.h"/>
<term file="dlib/geometry/vector_abstract.h.html" name="dpoint" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="dpoint" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="rotate_point" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="point_rotator" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="point_transform" include="dlib/geometry.h"/>
......@@ -424,6 +426,15 @@
<term file="linear_algebra.html" name="point_transform_projective" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="find_projective_transform" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="rotation_matrix" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="line" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="signed_distance_to_line" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="reverse" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="intersect" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="count_points_on_side_of_line" include="dlib/geometry.h"/>
<term file="linear_algebra.html" name="count_points_between_lines" include="dlib/geometry.h"/>
<term file="algorithms.html" name="running_stats" include="dlib/statistics.h"/>
<term file="algorithms.html" name="running_stats_decayed" include="dlib/statistics.h"/>
<term file="algorithms.html" name="running_scalar_covariance_decayed" include="dlib/statistics.h"/>
......@@ -1429,8 +1440,12 @@
<term file="imaging.html" name="get_pixel_intensity" include="dlib/pixel.h"/>
<term file="imaging.html" name="hysteresis_threshold" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="sobel_edge_detector" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="find_dark_lines" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="find_bright_lines" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="image_gradients" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="suppress_non_maximum_edges" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="normalize_image_gradients" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="remove_incoherent_edge_pixels" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="zero_border_pixels" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="binary_close" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="edge_orientation" include="dlib/image_transforms.h"/>
......@@ -1443,6 +1458,7 @@
<term file="imaging.html" name="segment_image" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="find_candidate_object_locations" include="dlib/image_transforms.h"/>
<term file="dlib/image_transforms/segment_image_abstract.h.html" name="remove_duplicates" include="dlib/image_transforms.h"/>
<term file="dlib/image_transforms/label_connected_blobs_abstract.h.html" name="neighbors_24" include="dlib/image_transforms.h"/>
<term file="dlib/image_transforms/label_connected_blobs_abstract.h.html" name="neighbors_8" include="dlib/image_transforms.h"/>
<term file="dlib/image_transforms/label_connected_blobs_abstract.h.html" name="neighbors_4" include="dlib/image_transforms.h"/>
<term file="dlib/image_transforms/label_connected_blobs_abstract.h.html" name="connected_if_both_not_zero" include="dlib/image_transforms.h"/>
......@@ -1542,6 +1558,8 @@
<term file="imaging.html" name="random_color_transform" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="hough_transform" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="skeleton" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="encode_8_pixel_neighbors" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="find_line_endpoints" include="dlib/image_transforms.h"/>
<term file="imaging.html" name="hessian_pyramid" include="dlib/image_keypoint.h"/>
<term file="imaging.html" name="compute_dominant_angle" include="dlib/image_keypoint.h"/>
<term file="imaging.html" name="draw_surf_points" include="dlib/image_keypoint/draw_surf_points.h"/>
......@@ -1591,6 +1609,7 @@
<term file="imaging.html" name="scan_image_movable_parts" include="dlib/image_processing.h"/>
<term file="imaging.html" name="find_points_above_thresh" include="dlib/image_processing.h"/>
<term file="imaging.html" name="find_peaks" include="dlib/image_processing.h"/>
<term file="dlib/image_processing/scan_fhog_pyramid_abstract.h.html" name="default_fhog_feature_extractor" include="dlib/image_processing.h"/>
<term file="dlib/image_processing/scan_fhog_pyramid_abstract.h.html" name="num_separable_filters" include="dlib/image_processing.h"/>
......@@ -1598,6 +1617,8 @@
<term file="imaging.html" name="evaluate_detectors" include="dlib/image_processing.h"/>
<term file="imaging.html" name="scan_fhog_pyramid" include="dlib/image_processing.h"/>
<term file="imaging.html" name="find_dark_keypoints" include="dlib/image_processing.h"/>
<term file="imaging.html" name="find_bright_keypoints" include="dlib/image_processing.h"/>
<term file="imaging.html" name="scan_image" include="dlib/image_processing.h"/>
<term file="imaging.html" name="scan_image_pyramid" include="dlib/image_processing.h"/>
<term file="imaging.html" name="scan_image_boxes" include="dlib/image_processing.h"/>
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment