3 o Modify d2::point::{neg,pos}inf() assertions to allow for isinf()
4 implementations returning any non-zero value for any infinity.
6 o Re-adjust exposure in Irani-Peleg even when no non-linear stage is
9 o Adjust the default pixel definition threshold to account for the
12 o Use a rudimentary, single-pass estimate-based certainty approach for
13 incremental rendering.
15 o When --cx is positive, use only certainty calculated from the simulated
16 scene for estimating tonal registration in Irani-Peleg passes.
18 o Remove --qr default setting, as certainty and range-extension now have
19 established general defaults.
21 o Change --cx to 1 by default. Change --exp-extend to be disabled by
22 default. Remove these flags from default settings.
24 o Update driver names to omit file type qualifiers, according to the
25 convention used in HJ Hornbeck's Nikon D50 device configuration patch.
27 o Add Nikon D50 device configuration. (HJ Hornbeck)
29 o Update default settings according to time and image quality test
30 results. fine:triangle:2, gauss:0.75, and combinations of these
31 filters were suggested by HJ Hornbeck.
33 o Revise incremental renderer to store unnormalized values, for better
34 handling of small weights, and eliminate the minimum contribution weight.
36 o Add inlining and pre-calculation to Irani-Peleg. (HJ Hornbeck)
38 o Add gaussian sampling filter. (HJ Hornbeck)
42 o Remove binaries and documentation sections from the readme, and
43 prune non-package documentation for source distributions.
45 o Make the --follow alignment option default.
47 o Add help text for argument scope.
49 o Modify rotational alignment search to rotate about the centroid
52 o Revert 3D code to experimental status due to poor handling of
55 o Modify 3D code for slightly better handling of real image sets.
57 V 0.8.5-pre3 2006-Nov-03
60 o Fix a potential nondeterminism bug due to undefined seed bits in the
61 random number generator. (Fixes certain alignment failures.)
64 o Modify default settings using Irani-Peleg rendering to use the
65 triangle:2 incremental rendering chain.
67 o Add dynamic monte carlo alignment, and make this default.
69 o Correct --exp-mult arg count. (Gregory Maxwell)
71 o Allow file-scoped use of --perturb-upper.
73 V 0.8.5-pre2 2006-Oct-25
75 o Modify d2::image::scale() to correctly handle offsets for
76 scale factors other than negative powers of two.
78 o Reformat most help text to fit a 79 column screen.
80 o Add option --profile to measure elapsed real time for 2D alignment
81 and rendering operations.
83 o Add support for POSIX threads to 2D alignment and rendering code.
84 (Adding multithreading support was first suggested by Jody (Joe)
85 Harris. Testing of multithreaded code on multiprocessor hardware was
86 done by Gregory Maxwell.)
88 V 0.8.5-pre1 2006-Oct-16
90 o Incorporate all existing ALE documentation within the source
91 tree, and license under GPL. Begin migration of documentation,
92 including package documentation files, to a more flexible format.
94 o Modify ui_tty for correct operation under Hurd and DragonFly
97 o Revise README for autotools build and installation.
99 o Add file 'TODO', containing a to-do list for future releases.
101 o Add AUTHORS file, in accordance with the autotools packaging
104 o Revise Changelog name to autotools-approved ChangeLog
106 o Add NEWS file containing mailing list and Freshmeat announcement text
107 for versions through 0.8.4. This file is required by autotools.
109 o Add options --f{ex,crop} for frame-coordinate exclusion regions.
111 o Add autotools-based configuration, build, and installation facilities,
112 replacing the earlier Makefile-based configuration and build process.
114 o Modify --mc to be a file-scoped option.
116 o Modify command-line interface to accept certain scope operators, and
117 to increase uniformity of syntax.
119 o Move input interface details, including option handling, messages,
120 and help information, to the ui/ subdirectory.
122 o In d3::scene, always use filler spaces when --pc all is specified.
126 o Fix build error noted by Gregory Maxwell and re-tag version.
128 o Add fine-grain UI updates for 3D operations when using ui=tty. Reduce
129 tty UI status update frequency to at most once per second.
131 o Modify --3dpx to affect scene geometry rather than rendered
132 pixels. In certain cases, this change should allow removal of objects
133 close to the camera, hence revealing objects further away.
135 o Add to the Makefile a HASH_MAP option that can be used to assign the
136 3D spatial info structure a hash map type instead of the standard STL
137 map type. As this is likely to break things, it is currently
138 undocumented. HASH_MAP=GNU seems to compile with g++ 4.0.x but
139 has not otherwise been tested.
141 o Add support for synthetic focus via the --focus option, including
142 focus with horizontal and vertical tilt. Adding simulated focus
143 effects, including tilted and swung focal planes, was suggested by
146 o Add support for filtering 3D output with --3d-filter. This
147 is now the default behavior. Add an option --3d-chain for selecting
148 the filter chain used for 3D output. (See --hl and --hq for details.)
150 o Fix dimension comparison bug in d3::space::traverse::get_next_split().
152 o Implement an experimental approach to 3D subspace candidate selection;
153 this seems to perform poorly in most cases due to parallax effects.
154 See the source code for details. (Option --subspace-traverse)
156 o Add adaptive weight decomposition to 3D occupancy algorithms.
158 o Move low-level code concerning scale and level-of-detail from
161 o Move code concerning the nature of space from d3::scene to
164 o Interpolate to generate smooth 3D depth output.
166 o Remove older 3D rendering code.
168 o Add decimation options --di-upper, --di-lower, and --do-try for
169 reducing the complexity of 3D scene geometry, hence reducing the
170 number of operations required for occupancy updating.
172 o Add a script 'transform' for applying an inverse projective
173 transformation. Recalculation of transformations was suggested by Rob
174 Stewart (ALE mailing list message, 2005-Dec-28).
176 o Add dates to the Changelog, using dates in comments in the download
177 page on the website for versions prior to 0.8.0, and using dates from
178 the darcs repository for versions 0.8.0 and later.
183 o Add option --3dpx to specify 3D regions to exclude during final
184 rendering, using a single depth for each pixel in each frame of
185 rendered output. (Evaluates at the depth used in depth-rendering.)
187 o Add option --pc to specify which pairwise comparisons should be
190 o Add --et option to specify encounter thresholds.
192 o Clean up d3::scene to remove triangle-mesh code.
194 o Add an option --def-nn to allow definition of pixels in output images
195 to take values from neighboring pixels within a certain radius, in
196 cases where they would not otherwise be defined.
198 o Modify representation of 3D subspaces so that refined subspaces
199 are roughly cubical, and to allow refinements at multiple
202 o Fix bugs in 3D color and occupancy updating caused by bounding
205 o Add option --pa to set the number of point candidates accepted
206 for each pair of cameras (pairwise ambiguity). The default
207 of 3 preserves existing default behavior.
209 o Add option --oui for determining occupancy update iterations.
210 The new default is 10, halving the previous default of 20.
212 o Add option --tcem for scoring point candidates by third-camera
213 views. The default of 0 preserves existing default behavior.
218 o Add magical file 'dump:' for listing color values numerically
221 o Transition from a subdivisible triangle model to a subdivisible
222 space model, using a rudimentary implementation of Hans Moravec's
225 o Add support for view and depth output from arbitrary cameras.
227 o Add clipping plane options for 3D modeling.
229 o Fix various bugs in control-point alignment reported by Gregory
235 o Modify --q1 to use median-value incremental rendering.
237 o Add support for median-value incremental and Irani-Peleg rendering.
238 Using a statistic other than the mean was suggested by Scott Scriven
239 as a means of eliminating non-uniform artifacts. Using the median
240 seems to improve performance on Scott Scriven's example sequences, as
241 well as on the scaled output of fine:sinc*lanc for sparse inputs.
243 o Move ale-psf-calibrate to a new scripts directory.
245 o Add support for using control points in 2D and 3D alignment, as well
246 as in 3D model construction. New options include '--gs points',
247 '--cpf-load=', '--cpp-upper=', '--cpp-lower=', '--va-upper=', and
250 o Fix enumerator compile bugs noted by Gregory Maxwell and Tom Holroyd
251 to cause compilation failure under GCC 3.4.
253 o Add average magnitude comments to transformation data file output.
255 o Add a color adjustment option --exp-mult=.
257 o For 3D scene reconstruction, use a new unified surface model in place
258 of the depth-based multi-model approach used in 0.8.0. The new model
259 is adjusted on the basis of edge, angle, and color measures.
264 o Add a flag --exp-meta-only which causes tonal registration to be
265 performed only on the basis of exposure metadata (e.g., data encoded
266 as EPPM extended attributes). This approach was suggested by Gregory
269 o Set image offsets in d2/render/combine.h. This allows Irani-Peleg
270 rendering to work with --crop when the default rendering chain
271 includes more than one rendering invariant.
273 o Perform Irani-Peleg updating separately by channel instead of
274 using vector operations. This allows divide-by-zero and other
275 conditions to be localized to particular channels.
277 o Modify the Irani-Peleg renderer to use estimate-based certainty. This
278 seems to improve noise characteristics significantly.
280 o Write the results of tonal registration as a comment in the
281 transformation data file.
283 o Add support for PPM with extended attributes (EPPM). Attributes
284 handled specially include black level, shutter speed, aperture, and
287 o Add experimental cost calculation routines to d3/scene.h.
289 o Modify euclidean alignment so that perturb-upper and perturb-lower
290 apply to rotational arclength.
292 o Fix an index variable bug in d2::image_bayer_ale_real::trigger().
294 o Modify d2::exposure one-sided certainty so that the 'one-sided'
295 condition only applies to bright (>50%) areas. Applying it to dark
296 areas has no clear benefit, and an experiment by Gregory Maxwell
297 demonstrates that doing so can exacerbate noise.
299 o Modify d2::align to use unscaled input images when the scale factor
300 is larger than 1.0, or when a scaled_filter interpolant is being used.
302 o Modify d2::image::scale() to allow scale factors in the interval
303 (0, 1), and modify the command-line interface and help text to reflect
304 this change. --scale= now accepts all finite positive values.
306 o Add user interface options --ui=stream and --ui=tty. The latter
307 includes more detailed status output when terminal width information
308 can be obtained (esp. when IOCTL=1 is set at compile time), and is the
311 o Move certain user interface elements to the source subdirectory ui/.
313 o Modify d2::render::ipc::_ip_frame_correct() so that exposure
314 re-estimation uses a ratio of weighted sums instead of a weighted
317 o Modify d2::align::set_exposure_ratio() so that alignment
318 weights and certainty values are taken into consideration.
320 o Add a --cw option to weight alignment error by certainty values,
321 as suggested by Gregory Maxwell. This allows the contributions of
322 pixels at maxval to be diminished, which can increase alignment
323 success in certain cases.
325 o Replace functions d2::incremental::{increase,decrease}_extents()
326 with new functions that take exclusion regions into account. The new
327 functions should also fix a boundary-calculation bug reported by
330 o Modify d2/render/ipc.h to handle exclusion regions more effectively.
332 o Add --crop argument, which excludes the spatial complement of a
333 specified area over a specified set of frames. Gregory Maxwell notes
334 several advantages of internal over external crops, including
335 preserving knowledge of position in the uncropped frame, which among
336 other benefits allows natural handling of Bayer patterns.
338 o Modify rx_parameters in d2/render.h to use the ale_pos data type
339 instead of int. This should allow exclusion regions to be calculated
345 o Add incremental 3D reconstruction output.
347 o Remove the reference to a non-existent web page in
348 unsupported::fornow().
350 o Allow device-specific bayer patterns to be overridden by the --bayer
353 o Add an option for explicitly setting the camera view angle. Make the
354 default view angle 43.7 degrees.
356 o Implement optimizations that reduce the memory footprint when 3D
359 o Modify the 3D cost reduction algorithm. The new algorithm seems to
360 offer better results than the old in cases tested thus far, and does
361 not require that spatial extension be enabled.
363 o Add support for the Canon 300D with a 50mm f/1.4 lens (specifying view
364 angle), as well as experimental support for the 50mm f/1.4 lens at 1.4
365 (specifying a linear PSF). Also add support for 50mm f/1.8 and 85mm
366 f/1.8 lenses, specifying view angles as specified in the chart at
367 http://www.acapixus.dk/photography/angle_of_view.htm
369 o Don't output final data when --psf-match is specified. Doing so
370 would be useless anyway, and apparently causes a segfault.
372 o For cases where --psf-match is specified, modify the exposure settings
373 for the final input image (i.e., reference image) so that default
374 exposure is used, and modify the bayer pattern settings so that no
375 bayer pattern is used. (The first of these modifications requires
376 changes in the way that the array of input exposures is stored.)
378 o Optionally (when OPTIMIZATIONS == 1) use the output of the default
379 rendering chain directly as the Irani-Peleg initial approximation,
380 instead of copying it.
382 o Add support for PSF convolution and scalar multiplication.
384 o Move all PSF parsing logic to a new file: d2/render/psf/psf_parse.h
386 o Modify help argument processing so that several help arguments can be
387 specified at a time. Change the usage text to reflect this.
389 o Fix concatenated help output; outputs from all help options are now
390 represented. (3D help had been omitted in 0.7.2.)
392 o Add a check to ensure that at most one --q* option is specified.
393 The issue of multiple --q* options being specified was raised on the
394 ALE mailing list by Viktors Rotanovs. As there is currently no
395 support for this, generating an error message is appropriate.
400 o Add cost-minimizing code to reconstruct 3D scenes from 2D inputs.
401 (This code is currently very experimental.)
403 o Add rudimentary device support for raw, linear files from the Canon
404 EOS 300D (Digital Rebel) SLR. CRW files may not be directly usable
405 under all configurations; in cases where they are not, Dave Coffin's
406 dcraw program, with arguments -d -4, will produce a raw, linear PPM
407 file. See http://www.cybercom.net/~dcoffin/dcraw/
409 o Add a new option --gs-mo to establish a minimum overlap area for
410 global searches. This may allow more of Derrick Brashear's historical
411 USGS contour map image sets to align successfully. (from
414 o Allow the scale factor to be explicitly set to one. (from
417 o Print version information to stdout instead of stderr. (from
420 o Add a rudimentary test suite. (from 0.7.1-patch2)
422 o Add support for --perturb-{upper,lower}=x%, where the value x% is
423 specified as a percentage of the smallest image dimension. Use of
424 percentages to specify bounds was suggested by Jeff Epler. (from
427 o In render::is_excluded(), add the offset before checking exclusion
428 regions. (from 0.7.1-patch1)
430 o Add code to shrink the accumulated image in d2/render/incremental.h,
431 so that superfluously large undefined areas are never present. This
432 may improve operation of '--gs inner' in some cases. (from
435 o In d2::align::_align(), rescale default_initial_alignment in parallel
436 with the current transformation, so that it can be used again, if
437 necessary. (from 0.7.1-patch1)
439 o Add experimental options --perturb-source and --perturb-output, which
440 allow some refinement of the manner in which perturbation occurs.
446 o Fix --visp segfault when not using last: filters.
448 o Add help mode --hA to display all help pages.
450 o Add circular point-spread function option "circle=<diameter>".
452 o Add support for the ov7620 sensor (raw, linear modes only). (Updated
453 from 0.7.0-patch1 to specify a linear PSF.)
455 o Add global search option --gs to fix an alignment problem reported by
456 Bert Douglas. (Also included in 0.7.0-patch1.)
458 o Change transformation::eu_{get,set,modify}() so that externally
459 visible values are interpreted in the context of transformations
460 between scaled images, while internal values represent transformations
461 between unscaled images. (Bugfix for 0.7.0. Also included in
464 o d2/tfile.h: Ensure that eu[] parameters are properly scaled when read
465 and written. (Bugfix for 0.7.0. Also included in 0.7.0-patch1.)
467 o Fix order-of-operations bug in d2::image::scale_by_half(). (Bugfix
468 for 0.7.0 and earlier versions. Also included in 0.7.0-patch1.)
470 o Scale alignment weights for reduced level-of-detail in d2::align().
471 (Also included in 0.7.0-patch1.)
473 o For reduced level-of-detail alignment reference images, filter
474 using alignment weights. (Also included in 0.7.0-patch1.)
476 o Add speed optimizations for filtering in the Euclidean case, but
477 leave these disabled, since they are not safe. (Also included
483 o Add support for algorithmic alignment weights. This requires
484 POSIX fork() and execlp(), so add a Makefile variable for POSIX.
486 o Add support for frequency-based alignment weights. Angelo Pesce
487 had initially suggested the related (or more general) idea of
488 edge-based alignment. Frequencies were chosen because of their
489 conceptual simplicity. This option requires that FFTW be linked.
491 o Add support for specifying per-pixel and per-channel alignment weights
492 by providing an additional weight-map image, as suggested by Bert
495 o Remove support for USM.
497 o Add multiple default settings: q0, q1, q2, qn, qr.
499 o Partition help into sections, and add a section for transformation
502 o Add a device type for the XVP610 camera in 640x480 mode.
504 o Modify the divisor in d2/render/ipc.h to ensure convergence when
505 using multiple varieties.
507 o Add experimental handling for Bayer patterns.
509 o Add experimental video stabilization options.
511 o Add options to control output range. By default, use the range of the
514 o For incremental rendering, add general filtering options. Drizzling
515 options are no longer supported as such, but the help output offers
516 examples of how to emulate old drizzling/merging techniques (roughly).
518 o Add options --fail-default and --fail-optimal to select the desired
519 failure mode. --fail-default uses the default alignment in the case
522 o Add option --ex, to support exclusion volumes. Region exclusion was
523 first suggested to the author by Dewayne VanHoozer as a means of
524 solving the 'closed eye problem' in group photography.
526 o Add a barrel distortion adjustment multiplier and rate options.
528 o In d2/transformation.h, add parameters for barrel distortion correction.
529 In d2/tfile.h, add new barrel distortion command 'B'. The approach
530 taken is based on that of Helmut Dersch's PanoTools. (i.e., a
531 polynomial in radial distance is used.)
533 o Increment the latest supported transformation file syntax to version
534 3. This version supports transformation of the original frame, as
535 well as other new features described above.
537 o Move d2/gpt.h to d2/transformation.h.
539 o In d2::align, add support for original frame transformations.
541 o In d2::align::diff() and d2::align::set_exposure_ratio(), treat cases
542 with and without extension identically.
544 o Report unrecognized transformation file commands as errors instead of
547 o In the transformation data file handler, report incompatible
548 transformation types as an error instead of as a warning.
553 o Add a generalized, but rudimentary, implementation of Steve Mann's
554 notion of 'certainty'. For more details, see
555 http://wearcam.org/comparametrics.pdf
557 o Make stdin filter height and width independent of each other.
559 o In the Irani-Peleg renderer, skip frames that fail to reach the match
562 o Convert some operations on pixels and points into vector syntax
563 through C++ operator overloading.
565 o Add command-line options to select plain or raw PNM/PPM formats. Remove
566 makefile variable PLAIN.
568 o Add command-line options to select 8/16 bits per channel. Remove
569 makefile variable BITS.
571 o Revise the device configuration xvp610_320x240 to use a new
572 filter, reflecting driver updates (see the relevant web page
573 http://auricle.dyndns.org/xvp610/ ) and support for non-linear
574 colorspace point-spread functions.
576 o In the Irani-Peleg renderer, use a discrete Fourier transform to
577 ensure that each (point-spread function, backprojection kernel) pair
578 satisfies condition (9) from theorem 4.1 in the source paper.
580 o Add a point-spread function rasterizer.
582 o Extend the --replace option to support drizzling.
584 o Add support for exposure registration (i.e., matching of exposure
585 between different frames).
587 o Change -Os to -O2 in the Makefile, since -Os no longer achieves speedup
588 in the most critical codepaths.
590 o Add command-line support for summation of point-spread functions.
592 o Add a polymorphic point-spread function type psf and convert all
593 external interfaces of ipc.h and usm.h (formerly hf-enhance.h) to use
596 o Fix memory leak in d2/render/combine.h:get_image(). (This fix is also
597 included in a patch for 0.5.1.)
599 o Modify the USM renderer to write output when incremental output is
602 o In the USM renderer, now only defined pixel values contribute to the
603 mask. Also, undefined values are no longer masked.
605 o Update e-mail address information.
607 o Modify the Irani-Peleg calibration script into a general point-spread
608 function calibration script, using a new built-in point-spread function
609 evaluation module to rate candidates.
611 o Irani-Peleg and USM (High-frequency enhancement) renderers now
612 use the same point-spread function configurations. New options --ips,
613 --usm, --lpsf, --nlpsf, and --device subsume and extend the
614 functionality provided by options --ip, --ipc, and --hf-enhance.
616 o Add support for convolutions in non-linear colorspace. This feature
617 is enabled and configured by the command-line option --nlpsf. (See
618 also the item above.)
620 o Linearize channel values within internal data structures so that most
621 ALE operations are performed on linear data. Steve Mann discusses the
622 use of linear representations in image processing, e.g. in his paper
623 http://wearcam.org/comparametrics.pdf . We assume a simple power-law
626 o Modify the image class to use a floating-point data type for
629 o Modify the README to remove information covered in the User Manual.
631 o Revise identifiers and comments in d2/render/ipc.h to better reflect
632 the terminology used in the source paper by Irani and Peleg.
634 o Consoldiate code for computing the position and dimensions of a pixel in a
635 new coordinate system, and use this new abstraction to reduce the code
636 size of d2/render/ipc.h and d2/render/drizzle.h (Irani-Peleg and
637 drizzling code, respectively).
639 o Change several instances of 'double', 'float', and 'my_real' to
640 specialized floating-point types 'ale_real', 'ale_pos', 'ale_sreal',
643 o Fix offset calculation in d2/render/ipc.h. This is necessary for
644 Irani-Peleg rendering of image mosaics extending upwards or to the
645 left of the original frame. (This fix is also included in a patch for
648 o Move classes treating snapshots as two-dimensional data to directory
649 'd2' and create an associated C++ namespace. We can't use '2d' to
650 label the namespace, because the C++ grammar won't allow it.
655 o Several of the changes below are based on a suggestion by Bert
656 Douglas regarding support for 16-bit-per-channel color. See the README
657 for more information about relevant compilation options.
659 o Add information to the README regarding new compilation options.
661 o Modify the makefile to accept a variable BITS, and differentiate between
662 values BITS=8 and BITS=16. The latter option causes the macro BITS_16
663 to be defined during compilation.
665 o Modify image_rw.h to handle 16-bit data when BITS_16 is defined. Also,
666 differentiate between 8-bit and 16-bit ImageMagick builds by checking
669 o In ppm.h, add carriage returns to the set of recognized whitespace
672 o Modify ppm.h to read 16-bit binary PPM files (added to the PPM format
675 o Modify ppm.h to read plain (ASCII) as well as raw (binary) PPM files.
677 o Add a Makefile option for writing plain (ASCII) PPM files instead of
678 raw (binary). This option causes the macro PPM_PLAIN to be defined
681 o Modify ppm.h to write plain (ASCII) PPM files when PPM_PLAIN is defined.
683 o Modify ppm.h to output 16-bit PPM files when BITS_16 is defined.
685 o Modify ppm.h to scale input intensity values based on the internal
686 representation depth (8 or 16 bits per channel).
688 o Add a file 'channel.h' to describe the color channel data type and
689 valid range, both being dependent on whether BITS_16 is defined. Use
690 these generic values, where appropriate, throughout the ALE codebase.
695 o Remove certain cloning operations from align.h, but since cursory tests
696 indicate that scaling with --scale is much more expensive than cloning,
697 allow cloning to occur when scaling occurs.
699 o Add more comment documentation to header files.
701 o Emit a more informative error message when the PPM file handler
702 encounters a file that is not a well-formed PPM binary file.
704 o Move rendering implementations to subdirectory 'render/'.
706 o Modify semantics of alignment following by applying to file-loaded
707 initial transformations a delta transformation D = AB^-1, where B is
708 the previous frame's (unmodified) initial transformation, A is the
709 previous frame's final transformation, and ^-1 is the inverse operator.
710 Using --follow with --trans-load can now improve alignment success when
711 working with panoramic image mosaics.
713 o We're not using simulated annealing, so remove comments that claim
716 o Remove certain unsafe assumptions from ipc/* (now render/ipc/*); this
717 change includes a modification to ipc.h (now render/ipc.h). Limits of
718 the rectangle used as an argument to the <type>::response() function
719 are now explicitly given rather than implied.
721 o Change documentation for --replace.
726 o Modify Irani-Peleg renderer to produce incremental output.
728 o Modify Monte Carlo alignment so that the same set of accumulated image
729 pixels is examined when comparing two transformations.
731 o Add an option for additional upper-bound constraint of rotational
732 perturbation. Jeff Epler had noted that the
733 perturb-upper/perturb-lower options alone may not be suited to working
734 with sets of higher-resolution images.
736 o Resolve a bug introduced in version 0.1.1 which had limited rotational
737 perturbation to 2 degrees.
739 o Add an option to adjust the maximum level of detail used. The default
740 maximum is 2/perturb, which matches the default for previous versions
743 o Modify names of temporary files used in the calibration script.
748 o Add a replacement option suggested by Jeff Treece. This option can
749 be used to facilitate video stabilization.
751 o Irani-Peleg: add support for configurable modules, including
752 a module that reads its configuration data from standard input
753 and a module calibrated for an IBM (Xirlink) XVP610 in 320x240
754 mode. A rudimentary script for calibrating devices with unknown
755 projection functions has been included as ipc/ale-calibrate.
757 o Irani-Peleg: don't truncate correction values in backprojection.
759 o Irani-Peleg: don't multiply the weight function by a scale correction
760 factor in backprojection.
762 o Rename render::operator()() to render::sync(). Update subclasses
768 o Allow --ip and --hf-enhance to be used together. [Modify
769 ip::operator() and hf_filter::operator() so that multiple
770 post-processing filters can be used.]
772 o Initialize the drizzled image to be uniformly black. This fixes
773 a bug introduced in 0.4.2.
775 o Allocate the correct size image for drizzling. This fixes a bug
781 o Ensure correct transformation scaling and difference calculation after
782 alignment is complete, even if alignment terminates with a reduced
783 level of detail (i.e. perturb-lower > 2).
785 o Separate the scale() function into two cases: scale() is reserved for
786 scale factors greater than or equal to one, as in user-specified
787 scaling; and scale_by_half() is used for the scale factor 0.5, needed
788 to provide multiple levels of detail in alignment. Modify the
789 scale_by_half() case to reflect source pixel values more fairly. (This
790 change affects alignment behavior at reduced levels of detail.)
792 o Reduce whitespace in usage() message.
794 o Modify align::diff() to use correctly scaled image offset values.
795 (Bugfix for alignment with --extend.)
797 o Modify align.h so that the image definition array is correctly scaled
798 for all levels of detail. (Bugfix for alignment with --extend.)
800 o Modify image.h to correctly handle image offset values. (Bugfix
801 for --extend operation.)
803 o Remove call to stat() in image_rw.h. This improves OS independence,
804 and apparently also allows URL-specification of images when compiled
805 with ImageMagick support.
807 o Modify image file handler interface so that input files are not scaled.
808 Perform scaling, when necessary, outside of the file handler. Since
809 unscaled data is needed by the drizzling and Irani-Peleg code, this
810 fixes bugs in those modules.
815 o Modify the Monte Carlo pixel selection algorithm so that the expected
816 number of sampled pixels more closely matches the specified number in
819 o Specify the binary mode flag to fopen() calls in the PPM file handler.
820 This change resolves a bug discovered by Benjamin Hill.
825 o Add support for Monte Carlo alignment of images, in which only a
826 random subset of pixels is sampled.
828 o Add an option to disable incremental image output.
833 o Implement an image reconstruction option based on the work of Irani and
836 o Clean up source; move rendering and alignment functions into separate
839 o Change valid drizzling diameter interval to (0,1].
841 o Modify PPM file reader to correctly handle comments in PPM files; this
842 resolves a bug pointed out by Benjamin Hill.
844 o Add missing typecast to resolve compile warning under GCC 3.2.
846 o Make changes to gpt.h that may reduce the Euclidean alignment time.
848 o Move ale.c to ale.cc to follow one of the many C++ naming conventions.
853 o Change language to C++.
855 o Code cleanup in ale.c, image.h, and gpt.h.
857 o Separate into a number of subsections the option descriptions in the
860 o Add URL information to README
862 o Add an experimental drizzling implementation. Research on drizzling
863 by Richard Hook and Andrew Fruchter was brought to the attention of
864 the author by Dan Kaminsky.
866 o Add an option to use the previous frame's final transformation as the
869 o Change pixel weight data type so that more than 255 images can
872 o Add support for increased spatial extents. Research on increased
873 spatial extents by Steve Mann was brought to the attention of the
876 o Modify the transformation algorithm to use transformations from the
877 domain of the input image to the range of the target image. This
878 change is reflected in transformation files. Maintain support for
879 old transformation file versions, however.
881 o Add version information to transformation files and update the version
885 V 0.4.0p1 2003-Apr-04
887 o Merge a bugfix from Debian ALE package version 0.4.0-1. Without this
888 fix, an assertion failure can occur when ImageMagick support is
889 enabled. Original fix implemented by Nick Rusnov.
894 o Add support for general scaling parameters (i.e. all scale
895 factors >= 1.0). --scale2, --scale4, and --scale8 are now deprecated,
896 and replaced with a more general --scale=x argument.
898 o Support enhancement of high frequencies even when the scale factor is
901 o Modify the high-frequency filter algorithm to support filtering near
902 the border of an image.
904 o Modify the high-frequency filter function so that the response is now
905 (4/n)^2 times what it used to be, where n is the scale factor.
910 o Don't filter high frequencies if we're not enhancing them. This
911 improves efficiency, but should not affect output.
917 o Output average match statistic.
919 o Update informational output to transformation files to include original
922 o Add support for post-enhancement of high-frequency details.
927 o Add support for loading and saving transformation data.
929 o Add an option to set the upper bound on correction steps.
931 o Modify Makefile to compile with size optimization, since this
932 appears to improve speed characteristics in some cases.
935 o Move transformation data structure and related functions to
936 a separate header file.
941 o Add support for general projective transformations, and change/add
944 '--translation' used to be '--no-rotate'
945 '--euclidean' used to be '--rotate'
946 '--projective' is a new option
951 o Add optional support for .jpg, .png, .tiff, and other file formats
952 through ImageMagick. Translation using ImageMagick was suggested by
956 o Don't segfault if a file can't be opened.
961 o Use reduced levels of detail in coarser alignment steps to increase
964 o Change perturbation step lower-bound check from less-than ('<') to
965 less-than-or-equal-to ('<='). Change default lower bound
966 to .125 from .1 (the behavior is the same, but the value
969 o Change perturbation initial value to 32 pixels from
970 16 pixels, since coarse alignment is now less expensive.
972 o Modify scale function so that the borders aren't
973 corrupted (too badly).
978 o Add support for correction of rotation about the lens axis.
980 o Add an optional match criterion for merging.
982 o Add an option to adjust minimum step-size.
984 o Add flags to specify which channels to use for alignment. Use
985 a summation of all channels by default.
987 o Add an option to modify the error metric exponent, and change
988 the default value from 1 to 2.
990 o Display information about how well a new image matches
991 the accumulated image.
993 o Add a rule for ale.exe to the Makefile.