* Added all standard morphological transformations
* Should handle grayscale dilation/erosion
* Added test cases and improved code structure
* Should handle multichannel images
* Added all standard morphological transformations
* Improved comments and some other things
* Applied adviced changes
* Applied adviced changes
* Should handle grayscale dilation/erosion
* Checking
* Added test cases and improved code structure
* Added command line control
* Added command line control
* Rectified some things
* Rectified some more things
* Improved comments
* Improved comments
* Improved doxygen comments and added more test cases
* Improved compatibility for builds and rectifying whitespace use
* Minor improvement in comments
* Did clang formatting
* pushed enum class inside namespace 'detail' and some other things
* Should handle multichannel images
* Clang formatting attempt
* got rid of if/else comparators for target_element
* Adds morphology.hpp declaration in boost/gil.hpp
* Fix newline
* (std::max)(a, b) instead of std::max(a, b)
* Improved Formatting
Support construction from step_size, step_count, and a function for angles
Implement angle and radious version of Hough line transform and adds a demo
with static line that goes over secondary diagonal.
Implement incremental line raster
Implement naive line raster
Implement Bresenham line raster
Leave only Bresenham line rasterization
Naive and incremental algorithms were removed because they are supposed
to produce the same results anyway.
The reason for diverging results is inaccuracy of floating point numbers
Add circle rendering through trigonometric functions, using
arctan(1 / (radius + 1)) as minimal angle step.
Trigonometric circle rasterizer does not follow circle equation, but still
produces very round shapes.
A new testing methodology needs to be devised for this rasterizer.
The new version accepts start and points inclusively and tries to use
canonic representation during computations.
Slope decided to be is (diff_y + 1) / (diff_x + 1).
Add basic GitHub Actions configuration based on mp11
Remove Actions jobs using GCC 4.7 and 4.8 - unsupported compilers
Run b2 with --abbreviate-paths on Windows
The -std=c++1z is broken for clang-4.0 but no need to test it
Add -mbig-obj to GCC on Windows
- That is to avoid string table overflow and file too big
Define _GLIBCXX_USE_CXX11_ABI=0 for clang 3.5, 3.6, 3.7
- Should help avoid linker error:
`undefined reference to std::ios_base::failure::failure(char const*, std::error_code const&)`
Disable certain check in algorithm_channel_relation test for clang<3.8
A new histogram class proposed with close suport for gil
image constructs.
Shift the stl support implmentation to extension to
serve as example for overloading fill_histogram.
Add cumulative histogram and histogram normalization.
Co-authored-by: debabrata1 <debabrata@goodhealthapp.com>
The output type must be floating point, thus a check was added to make sure it
is the case. Though I had to add specific cases for float32_t as std::is_floating_point does not
consider it a floating point type
The accumulate part was wrong, it multiplied by delta_t on every sum, which is wrong
Use 8 way nabla compute.
This is just different discretization of Laplace operator
https://en.wikipedia.org/wiki/Discrete_Laplace_operator
Laplace stencils are now the same as in mathematical notation
The new function will provide a uniform way to generate stencils
by making sure directions are indexed properly
Add only required stencil points:
The 5 points Laplace stencil is now adding only required points and not assuming that others are zero
* Fix conversion of rgb to signed cmyk(boostorg#479)
* changed naming of dst_us_t to uint_t and undid the formatting change
* test for conversion of rgb to cmyk, PR #522.
* small formatting changes/fixes
* Build configuration update for PR #522
* removed unused header file
Correct calculation to correctly map CMYK 0 to minimum
and 1 to maximum of input and output channel types.
If float-point division is necessary, use double instead of float
which may be too narrow for large operands. cmyk32_pixel_t is based on
uint32_t which multiplication by its maximum may yield result too
large to fit 32-bit float.
For example, (uint32_t(c) - uint32_t(k)) * float(s) for
c = 4294967295, k = 0, s = 1.0
results in 4294967300 value which does not fit uint32_t.
Fixes#406, but does not fix#479
Since we do not require C++17 yet with its `if constexpr
that would discard instantiation of allocator's `operator=`
when the statement is not active, we have to use the tag
dispatching with appropriate used of
`propagate_on_container_move_assignment` and
rebound `allocator_type` to choose the POCMA.
Minor coding style fixes.
Improve utilities sending channels, pixels and other GIL objects to
output stream for logging of test errors.
Fix missing namespace qualification in IO tests.
Comment TARGA test case with FIXME that was accidentally
uncommented in cc64bdd1a4
Motivation is to:
- use on simpler and light test framework,
- eliminate dependency on libraries like Boost.MPL,
- achieve faster compilation times for CI builds (20% seems feasible)
- have test programs easy to run and debug
- avoid macros
Remove outdated FIXME-s for bugs that have been already fixed.
Fix off-by-one bug in test/core/test_fixture.hpp generators.
Minor corrections and tidying up.
Add missing test assertions to numeric extension tests.
Fixes#458
* Add tests for image_view locator and iterators
Update follows PR #450 extending tests coverage and
still includes test case for fix of issue #432
* Comment UB if specified increment advances non-incrementable iterator
Add TODO comments asking about clarification of valid pixel range specification
for end iterators, especially of empty view (image).
Since `std::uninitialized_fill` performs a placement new on each
element in the range, it is important to ensure core and packed
pixel types behave correctly.
Split general purpose test fixtures into core/test_fixture.hpp.
Refactoring to unify use of the common framework across GIL tests.
Merge `unit_test_utility.hpp` into `unit_test.hpp` for convenience.
Add several test cases to `gil::point` tests.
* Implement Sobel and Scharr operators
This commit adds Sobel and Scharr
operators with support for 0th and 1st
degrees with other degrees planned for
later
* Migrate and fix Harris example
Generate Harris entries now uses
signed image view.
The Harris corner detector example
now uses the Scharr filter generator
and convolve_2d to reduce amount
of code needed.
* Fix and migrate Hessian example
The Hessian example now uses signed
image views and uses newly added kernel
generators to compute gradients
* Fix Harris and Hessian tests
The tests broke due to migration to
signed views in algorithms, but tests
were not adjusted
* Fix Jamfile for example/sobel_scharr.cpp
* Cosmetic changes
* Commented out fail tests
* Fixed pixel16 used in image16s
In Harris and Hessian tests, unsigned
pixel values was used to construct
signed image, which was causing
appveyor to error out.
* Reenable failing targets
* Unify kernel generator interface
This commit makes all kernel
generator functions to return kernel_2d
and adapts dependant threshold
function to use the new interface
* Migrate Hessian and Harris tests
Migrate Hessian and Harris tests to new
interface for kernel generators
* Migrate Harris and Hessian examples
Harris and Hessian examples now use
new interface for kernel generation
* Migrate simple_kernels tests
simple_kernels are now using kernel_2d
interface
* Add missing return
Normalized mean generation had missing
return at the end of the function
* Adapt code to namespace move
This commit reacts to kernel_2d,
convolve_2d being moved to
namespace detail
* Hessian plain determinant
This commit partially implements
Hessian corner detector, but only
uses determinant as Hessian response
* Implement full Hessian corner detector
This commit complements last one
by summing in a window and applying
det - k * trace * trace formula to
final Hessian response
* Add docs and make code align with docs
This commit adds docs to new functions,
and makes function and variable names
align with docs to not confuse readers
* Use determinant as response function
A-KAZE uses only determinant in it's
response, and since for now Hessian
is only a mean to advance A-KAZE
implementation, response function is
adjusted to use only determinant
* Create simple test for Hessian detector
This commit adds an s at the end of
function name to make it uniform with
another detector, and adds a simple
test for sanity check
* Improve documentations for d params
dx and similar naming seems to be
confusing, improved documentation
to explain the naming convention
* Address minor comments about style
* Address type based issues
Mostly changes to constness and
integral types, with small cosmetic
changes mixed in
* Fix typo and address review comment
Fixes typo in call for Hessian and
addresses a review comment about
replacing multiple exact indexing
calls to one with a reference
* Reorder includes in Hessian example
* Address review comments
Add literature reference to luminosity
computation, and perform some cosmetic
changes.
* Restore previous state of Harris
This is a fresh start of branch
Harris with manual copying and
placement of code from previous
versions
* Use supplied weights in Harris
Weights passed as arguments were not
used prior to this commit
* Address review comments
Mostly moving brackets and adding
const, but also address MSVC's
max macro
* Make namespace qualification consistent
In file numeric.hpp, there is
full qualification, but harris.hpp
didn't have full qualification,
thus full qualification is added to
harris.hpp
* Add copyright and license notice
Add static assertions to verify channel types are convertible to
expected result type.
Add new and update existing Doxygen comments describing interfaces in
`extension/numeric/channel_numeric_operations.hpp`.
* Implement generation of simple kernels
This commits implements mean and
Gaussian kernel generators,
roughly documents them, and has
tests. But tests for Gaussian have
accuracy problem
* Use difference for testing Gaussian
This commit applies percentage based
difference check on generation of
Gaussian kernels, as exact values are
dependant on platform. The error margin
is set to 5% in this commit
* Remove debugging code
In simple_kernels.cpp, there was
some debugging code, mostly streams
related stuff and commented out code,
which is now removed.
* Address review comments
This commit adds inline to kernel
generator functions, uses
std::ptrdiff_t instead of long int,
resolves some conversion warnings
and does small cosmetic changes
* Remove useless iostream include
There was stray include from debugging
code, now removed
* Use \code and \endcode for formula
This commit applies formatting to
formula of normalized mean
* Change are_equal to is_equal
Change a function name in
simple_kernels.cpp test
* Remove redundant function
In simple_kernels.cpp test there was
a function that was not used at all,
now removed
* Cosmetic changes
Mostly opening brackets moved
