Age | Commit message (Collapse) | Author | |
---|---|---|---|
2015-10-15 | fix corner-case issues when rescalingrescaling | Ronan Collobert | |
2015-09-16 | Merge pull request #106 from bshillingford/master | Soumith Chintala | |
Fix JPEG magic number to handle more JPEGs | |||
2015-09-16 | Fix JPEG magic number to handle more JPEGs | Brendan Shillingford | |
for example, these are all valid: FF D8 FF E0 FF D8 FF E1 FF D8 FF E3 FF D8 FF E8 FF D8 FF DB sources: https://en.wikipedia.org/wiki/List_of_file_signatures http://filesignatures.net/?page=all&order=SIGNATURE&alpha=J | |||
2015-09-14 | Merge pull request #105 from Moodstocks/warp_bounds | Soumith Chintala | |
warp: fix bounds checking for bicubic mode | |||
2015-09-13 | warp: fix bounds checking for bicubic mode | Cédric Deltheil | |
This is related to commit b5fc8d0. | |||
2015-09-12 | Merge pull request #104 from colesbury/warp | Soumith Chintala | |
Fix bicubic interpolation in image.warp | |||
2015-09-12 | Fix bicubic interpolation in image.warp | Sam Gross | |
The existing bicubic sampling produces artifacts. Fixes issue #103 | |||
2015-09-04 | Merge pull request #102 from etrulls/master | Soumith Chintala | |
Fix boundary/quantization artefacts on rotation. | |||
2015-09-03 | Fix boundary/quantization artefacts on rotation. | Eduard Trulls | |
2015-09-02 | Merge pull request #99 from nicholas-leonard/rtd | Soumith Chintala | |
readthedocs | |||
2015-08-30 | Merge pull request #101 from Moodstocks/png_generic | Soumith Chintala | |
png.c: move apart non generic functions | |||
2015-08-30 | png.c: move apart non generic functions | Cédric Deltheil | |
2015-08-29 | Merge pull request #100 from Moodstocks/png_error | Soumith Chintala | |
png.c: bubble up error messages from libpng | |||
2015-08-29 | png.c: bubble up error messages from libpng | Cédric Deltheil | |
2015-08-29 | Merge pull request #95 from edward-zhu/pngfix | Soumith Chintala | |
fixed open 1-bit png on Mac | |||
2015-08-27 | readthedocs | nicholas-leonard | |
2015-08-25 | Merge pull request #98 from Aysegul/err_messages | Soumith Chintala | |
add in-place error message for rotate | |||
2015-08-25 | add in-place error message for rotate | Aysegul | |
2015-08-24 | Merge pull request #96 from colesbury/saturate | Soumith Chintala | |
Fix clampImage function call to saturate | |||
2015-08-24 | Fix clampImage function call to saturate | Sam Gross | |
2015-08-21 | fixed open 1-bit png on Mac | Edward Zhu | |
2015-08-17 | Merge pull request #93 from colesbury/byte | Soumith Chintala | |
Support saving torch.ByteTensor images | |||
2015-08-17 | Support saving torch.ByteTensor images | Sam Gross | |
2015-08-14 | Merge pull request #92 from gheinrich/dev/issue_91 | Soumith Chintala | |
Fix call to processPNG from decompressPNG - closes #91 | |||
2015-08-14 | Fix call to processPNG from decompressPNG - closes #91 | Greg Heinrich | |
2015-08-14 | Merge pull request #90 from jonathantompson/warp_clamp_val | Soumith Chintala | |
Added user specified pad value to warp function. | |||
2015-08-14 | Added user specified pad value to warp function. | Jonathan Tompson | |
2015-07-31 | Merge pull request #89 from torch/bilinearfix | Soumith Chintala | |
fixing scale bugs in bilinear (src = 1) and bicubic (dest = 1) | |||
2015-07-31 | fixing scale bugs in bilinear (src = 1) and bicubic (dest = 1)bilinearfix | soumith | |
2015-07-30 | Merge pull request #86 from jagapiou/bicubic | Soumith Chintala | |
Add bicubic interpolation to image.scale. | |||
2015-07-30 | Merge pull request #88 from samehkhamis/master | Soumith Chintala | |
Compiling with libpng and libjpeg under Windows | |||
2015-07-30 | Compiling with libpng and libjpeg under Windows | samehkhamis | |
2015-07-24 | Merge pull request #87 from samehkhamis/master | Soumith Chintala | |
Making things work under windows | |||
2015-07-24 | Work under windows | samehkhamis | |
2015-07-22 | Add bicubic interpolation to image.scale. | John Agapiou | |
* Added 'bicubic' method to image.scale. * Added unit tests of scaling. Does interpolation for both upsizing and downsizing and does not preserve the average pixel intensity. Note this is different from 'bilinear' which preserves the average pixel intensity on downsizing but not on upsizing. At edges the missing data is added by extending the first/last line segment. | |||
2015-07-13 | Merge pull request #83 from mtourne/fix_pgma | Soumith Chintala | |
Fixing broken PGMA loading - Adding test | |||
2015-07-13 | Fixing broken PGMA loading - Adding test | Matthieu Tourne | |
2015-07-09 | Merge pull request #70 from colesbury/lua52 | Soumith Chintala | |
Support Lua 5.2 | |||
2015-06-26 | Support Lua 5.2 | Sam Gross | |
2015-06-19 | Merge pull request #76 from krasin/two-args-png-load | Soumith Chintala | |
png: add support for 16-bit PNG files. | |||
2015-06-18 | Merge pull request #78 from lukemarris/pyramid | Soumith Chintala | |
image.gaussianpyramid non-default type input fix | |||
2015-06-18 | Added a primitive test file for gaussianpyramid | marris | |
2015-06-18 | Fix for using non-torch.DoubleTensor inputs to image.gaussianpyramid | marris | |
2015-06-14 | Fix 16-bit color PNG; add tests for color PNG files; use inline tensor ↵ | Ivan Krasin | |
expectations. | |||
2015-06-14 | Merge pull request #72 from Maratyszcza/patch-1 | Soumith Chintala | |
Avoid recomputing sin/cos in rotate main loop | |||
2015-06-12 | Delete unused code | Ivan Krasin | |
2015-06-12 | Add test_png.lua that checks gray images (8- and 16-bit) | Ivan Krasin | |
2015-06-07 | png: add support for 16-bit PNG files. | Ivan Krasin | |
Note: libpng.load now returns two arguments. The first argument is the result tensor, and the second argument is the color bitness. It will be either 8 or 16. For images with smaller color bit depth (for example, 4), the value is upgraded to 8 bits. | |||
2015-06-02 | Use sinf and cosf instead of non-standard sincosf | Marat Dukhan | |
2015-06-02 | Avoid recomputing sin/cos in rotate main loop | Marat Dukhan | |