Git Download

The most up-to-date SIP version can be downloaded from Git. These instructions should work for any Unix, althought they have been tested on Ubuntu Linux and Mac OS. For Ubuntu users, there is an automation step below.

wget -O- https://sourceforge.net/p/siptoolbox/siptoolbox/ci/master/tree/install-sip?format=raw | sh

Additional instructions, including a web view of the repository can be found at [Souceforge generic instructions].

Please keep the following notes in mind when dealing with the SIP development version.

• SIP is beta software, and consider the development version as alpha software: it is intended for developers and advanced users, it is not stable, and will change very fast. Excuse us for programming errors, which may be frequent. Please contribute with corrections, suggestions, and code.
• Please read the *.txt files that come with the SIP development source, before any serious attempt to use or modify it. They have many information that are not in this home page.
• There might be some personal comments throughout the code.

The following text (in Portuguese) has information on SIP's design and internals. Appendix B shows step-by-step how to add a new routine to SIP.

Old windows port

There is a new complete tasklist at Sourceforge:

TO-DO List

To see where you can contribute, browse the open tasks which are unassigned. Please don't hesitate in suggesting other tasks.

We need somebody to port SIP for windows.

I've been working a lot with distance transforms and related entities such as skeletonization, Voronoi diagrams, and level sets, since these are the subject of my grad research.

Another important addition currently under development is the use of SciLab integer types. This will make SIP much more efficient.

Making a true-size fast image display is one of the most important improvements to be made. Image display is still not good in SIP/SciLab, because Matplot uses rectangle primitives instead of pixel buffer X11 primitive. Please help us!

Many other additions and improvements need to be made. See the to-do list above.

Some years ago, SIP had only a single developer - me. There are very few (if any) software that got robust with a team of a single person. Free, open-source software relies on joint development to be as good as commercial software. ImageMagick has about 5 developers, GIMP has about 100, and KDE about 1000. I try to make SIP as useful as I can, but cannot make any miracle alone. A single person developing free software is not like a company which is paid for doing what people want. As such, I have my personal interests with SIP, and must give end-user needs a lesser priority. (In the last months this has been changing. We now have three active co-developers contributing with code and testing: Jocelyn Druel, Leandro Estrozi and Bruno Pincon).

My primary objective with SIP is for learning. I want to improve my knowledge of computer vision algorithms and some C programming, while having some fun making a useful package. Making a lightning-fast, rock-solid and shiny package is not my direct objective. But this does not mean that SIP cannot be all that. It just means that I personally will not be much concerned with those issues.

SIP is free software, and as such I invite you to help me develop it. If you need a very fast package with very easy installation procedure, you are invited to contribute. If you need a particular image processing routine, please write it and send it to me. It's an advantage you get being able to directly shape SIP to your particular needs. In addition, if your contribution is continued, I am willing to share the authorship with you. For my part, I'll be concentrated in the algorithms I use for my particular research and interests. Of course, I will be accepting suggestions and feature requests, but I will give priority to the stuff I have personal interest.

The versioning of the SIP stable branch follows the overall GNU recommendation. The format is Major.Minor.Patchlevel. However, the interpretation of the numbers is not standard, so for SIP it is described in the table bellow.

Number	Why it changes	Period of Change (ideally)
Major	The software goes to a totally new level of stability and functionality	1 - 2 years
Minor	Significant functionality has been added	1 - 2 months
Patchlevel	Bugfixes and minor improvement	1 - 2 weeks

When the Major number is zero, we consider that SIP is in a special evolving stage. In that stage, changes and improvements in occur at a fast pace. The software is useful and fairly stable, but has not reached the final objective. In the case of SIP, the final objective is to make a complete image processing package, with functions for every task in this area: image I/O, segmentation, grayscale image processing, binary image processing, shape classification and recognition, etc.