Brian Popow
|
b842c07fa8
|
Fix issue with white is zero with 24 and 32 bit
|
5 years ago |
Brian Popow
|
b29581e57a
|
Add support for decoding tiff's with 32bit float gray pixel data with min is white
|
5 years ago |
Brian Popow
|
9d6b7a6204
|
Add support for decoding tiff's with 32bit float gray pixel data with min is black
|
5 years ago |
Brian Popow
|
3d02cfdbf0
|
Add support for decoding tiff's with 32bit float rgb pixel data
|
5 years ago |
Brian Popow
|
9b1276d2da
|
Add test images for miniswhite 32 bit
|
5 years ago |
Brian Popow
|
8f1e43a95d
|
Additional test images
|
5 years ago |
Brian Popow
|
8c7ee589e6
|
Add support for decoding 32bit per channel color tiff with planar pixel data
|
5 years ago |
Brian Popow
|
f118b81174
|
Fix issue calculating the stripIndex for planar tiff's
|
5 years ago |
Brian Popow
|
9c585c4899
|
Add support for decoding 32bit per channel color tiff with contiguous pixel data
|
5 years ago |
Brian Popow
|
c34a1ed024
|
Add support for decoding gray 32 bit tiff's
|
5 years ago |
Brian Popow
|
e700b972fd
|
Add support for decoding gray 24 bit tiff's
|
5 years ago |
Brian Popow
|
28c1356a76
|
Add support for decoding 24bit per channel color tiff with planar pixel data
|
5 years ago |
Brian Popow
|
6e9cff93f4
|
Add support for decoding 24bit per channel color tiff with contiguous pixel data
|
5 years ago |
Brian Popow
|
f868b7a2aa
|
Add min is white 16 bit gray big endian test file
|
5 years ago |
Brian Popow
|
9e12a72a55
|
Fix loop bounds for little endian, add little endian test file
|
5 years ago |
Brian Popow
|
a1ee0d638d
|
Decoding 16bit rgb planar now also respects byte order
|
5 years ago |
Brian Popow
|
cc16677172
|
Tiff decoder now respects byte order for 16 bit gray images with white is zero
|
5 years ago |
Brian Popow
|
5d888bef8f
|
Tiff decoder now respects byte order for 16 bit gray images
|
5 years ago |
Brian Popow
|
a7d44a435c
|
Use same byte order as IFD directory to decode pixels for 16 bit per channel data, fixes #1716
|
5 years ago |
Dmitry Pentin
|
99b24fefd3
|
Added images to quality estimation tests
|
5 years ago |
Dmitry Pentin
|
0ace1a042a
|
Added issue-1693 images & tests cases - all passing after fix
|
5 years ago |
Dmitry Pentin
|
2e5b0ad74d
|
Fixed baseline image invalid reference output png image
|
5 years ago |
Brian Popow
|
347279c258
|
Clamp color map index, fixes issue #1668
|
5 years ago |
James Jackson-South
|
2ab611fd5c
|
Use 5 bits for each component
|
5 years ago |
James Jackson-South
|
e773295622
|
Update refs to match new output
|
5 years ago |
James Jackson-South
|
78e0ab8181
|
Remove parallel processing & update refs
|
5 years ago |
Brian Popow
|
6281743b3b
|
Add support decoding for 16 bits per channel tiff's
|
5 years ago |
Brian Popow
|
bbd71e2ce7
|
Add support decoding for 12 bits per channel tiff's
|
5 years ago |
Brian Popow
|
3b18d705e3
|
Additional tests for gray tiff images
|
5 years ago |
Brian Popow
|
deed748525
|
Add support for decoding 10 bit per channel rgb tiff's
|
5 years ago |
Brian Popow
|
8e6fad805c
|
Add support for decoding 30 bit per pixel tiff's
|
5 years ago |
Brian Popow
|
42d5d9ee91
|
Add support for decoding 6 bit per pixel tiff's
|
5 years ago |
Brian Popow
|
c6f5a8aaa0
|
Add support for decoding 12 bits per pixel tiff's
|
5 years ago |
Brian Popow
|
881bb51f21
|
Make sure encoding 4bit paletted tiff rows are byte aligned
|
5 years ago |
James Jackson-South
|
a8d269f35a
|
Better test
|
5 years ago |
James Jackson-South
|
5ceba7116c
|
Ensure cdfMinSpan is cleared before use.
|
5 years ago |
James Jackson-South
|
dc0982f186
|
Faster Linear Transforms (#1591)
* Attempt to use same weight generation algorithm as resize.
* tests pass
* Identical output
* Update LinearTransformKernelFactory{TResampler}.cs
* Use new low allocation iterator
* Migrate projective transforms.
* Optimizations
* Smaller kernel
* Fix sampling accuracy
* Finalize and update refs
* Revert unnecessary changes
* Remove enumerator
* Actually save output for debugging.
* Use custom test png encoder for reduced memory environments
* Convolution should use scaled vectors
* Update TestEnvironmentTests.cs
* Try using doubles
* Moar double precision
* Fix radius calculation
* Test if issue is SIMD related.
* Detect runtime to determine pipeline.
* Fix stack overflow
* fix condition
* Try simplified scalar run
* Simplify unpremultiply scalar
* Update Numerics.cs
* Fix runtime environment
* Update ImageSharp.csproj
* Duplicate the caller with scalar versions
* Update method name, exclude from coverage.
* Don't save output during coverage tests for perf.
* Update src/ImageSharp/Common/Helpers/RuntimeEnvironment.cs
Co-authored-by: Anton Firszov <antonfir@gmail.com>
Co-authored-by: Anton Firszov <antonfir@gmail.com>
|
5 years ago |
James Jackson-South
|
a54bce1faa
|
Update refs.
|
5 years ago |
Brian Popow
|
1dbe583824
|
Allow encoding 4bit color palette images
|
5 years ago |
James Jackson-South
|
cfd123239d
|
Ensure LFS
|
5 years ago |
Franz Häring
|
2d3ab4ebd5
|
New overloads for binary threshold operations. The new argument ColorComponent defines the scalar color component to be used for threshold comparison: Luminance, Saturation or MaxChroma. Luminance is default and identical to previous versions. Saturation is the HSL saturation component. MaxChroma is calculated as the maximum of YCbCr chroma value, i.e. Cb and Cr distance from achromatic value. Background: This component shall discriminate colorful parts from achromatic parts in human perception. Very dark pixels, which are perceived as near black, can have high HSL saturation values if e.g. (rgb)==(4,0,0); this would definitely not be perceived as colorful by a human. The MaxChroma component will calculate them low.
|
5 years ago |
James Jackson-South
|
d183091ac4
|
Add output reference images
|
5 years ago |
James Jackson-South
|
ef1ee4d33b
|
Removed submodule
|
5 years ago |
Ildar Khayrutdinov
|
785bbc9277
|
Remove excess benchmark reports
|
5 years ago |
Ildar Khayrutdinov
|
ace0d184b0
|
Update benchmark results
|
5 years ago |
Ildar Khayrutdinov
|
48796f8241
|
Add benchmarks results
|
5 years ago |
Brian Popow
|
f0670af045
|
Handle TIFF image with incorrect IPTC data type (long instead of byte)
|
5 years ago |
James Jackson-South
|
f1a8e6e763
|
Santize descriptor bounds. Fixes #1530
|
5 years ago |
Brian Popow
|
7d51106880
|
Tiff decoder can now handle special case, when rowsPerStrip has the default value
|
5 years ago |
Brian Popow
|
0e5938f92a
|
Parse T4 Options: honor eol padding, if present
|
5 years ago |