Bonum Certa Men Certa

Software is Mathematics

Block diagram Depiction of the framework for detection and classification of 3-D data such as faces or internal organs, as well as benchmarking

Summary: Personal take on why software is just symbiosis of maths

This post is a concise summary of something I have been working on (source code will be uploaded at a later date when I tidy everything up). The figure at the top shows a breakdown of our existing framework, which basically depicts a program as a set of black boxes. Each box merely performs mathematical operations on volumetric matrices. It's maths. It can also be specified most precisely using equations (no need for explanation using code or pseudo-code). For the sake of simplicity, the block diagram contains only core components that are used irrespective of the approach tested. The file loaders, for example, are shown separately. They work very well and can elegantly load datasets based on a data selector. With the exception of test sets that are small (remnants of debugging), there are 6 families of data, some grouped in pairs, some grouped by training/target, some our lab's data in isolation, others for FRGC data from NIST. There are also correct and incorrect matches in isolation. These simplify the plotting of ROC curves in a largely streamlined fashion. The goal is the program is to test new metrics that can be used to analyse anything elastic such as tissue. It has uses in cardiac analysis, brains, and even faces (I have dealt with each of these data types). For large quantiles of 3-D data, about 70 GB of face data gets used.



The nose-finding part may as well be treated as a component that provides orientation o a form of segmentation (it can be a face or even an internal organ which we wish to model and perform binary diagnosis on). Depending on the datasets, different methods are used. Commonly, FRGC data is better off interpreted by finding nearest point, excepting noise. For the lab's data, it is preferable to choose the nearest point within a specified region (usually around the centre, no weighting/scoring based on location although that too would work). This can also be done using ICP, as described later (settings inherited from another box/module) or a Viola-Jones approach with face templates for training, although it is only partly implemented so far. Sphere intersection with plane, as per Mian et al. (with separate slider for radius), is another existing option, but it does not appear to outperform the simpler methods, which work most of the time given some reasonable boundaries (e.g. boundaries to dodge the hair region).

Having identified the tip of the nose correctly, we are cropping out what is left for rigid areas to be isolated. It is quite customisable. Various separation methods and boundary types like circle, ellipsoid, and rectangle have been tested, where circle is the most commonly used one that works in conjunction with binary masks. These come with many sliders and use measurements in X and Y to estimate real physical distances and then factorise pixel-space units, accordingly. There is also a slider for further manual tweaking. And still, it's all maths.

There are some other bits of operation that are worth mentioning; left out from the diagram in order to reduce clutter are smoothers, hole removers, outlier eliminators, and rounding up of values, all of which are optional and very much depend on the data at hand and how it ought to be treated. For instance, FRGC data hardly requires any smoothing. Lab data has offsets that need to be handled systematically depending on the image number. In fact, both datasets do need a lot of branching/forking in the code as their handling and even their size varies (the program is built to handle any image side with any aspect ratio, but for sub-regions to be defined it uses absolute and not relative coordinate inputs).

Then we come to the key part, which actually does more to contribute toward similarity measures. ICP becomes very important in case the initial alignment of the noses is deemed incorrect or the faces are tilted. In practice, assuming the faces are forward-looking and bend neither to the sides or top/bottom, ICP is not supposed to change much. The methods already available are Mian's early ICP method, Mian's most recent ICP method, Raviv's ICP implementation from 2008, and Raviv/Rosman ICP implementation from recent months or years. The program optionally applies translation and optionally it applies the rotation too. In many cases this does not seem necessary as ICP hardly modifies anything substantial.

The model part is not included in the diagram as there are many different things are can be done with a model. PCA, model-building, model assessment, file loaders for models (about 2 gigabytes for some), in addition to more basic measures on which assessment is applied, are basically all sorts of comparators which yield one value for each pair, then proceeding to the plotting of ROC curves (mostly automated following experimental design).

All the above is just mathematics. It can all be described using equations. To patent such stuff would be to claim a monopoly on equations, which means the monopoly covers a wide range (infinite even) of implementations. How can anybody defend the argument that software is not maths? Or that "innovative" software is somehow the exception? If many equations are already patented, how is one supposed to code safely? How can existing methods be enhanced without a violation?

Comments

Recent Techrights' Posts

Michael “Monty” Widenius: It Started in 1983 With Richard Stallman (RMS)
The other co-founder of MySQL is a bit notorious for confronting RMS rather viciously
For the Second Time in a Few Weeks Microsoft Lunduke Makes False Accusations Against Senior Red Hat Staff to Incite a Despicable 'Troll Army'
Nothing that Microsoft Lunduke claims of says can be trusted
su lisa && rm -rf /home/ibm/power
Novell was ruined by another person from IBM, Ronald Hovsepian
A Record Demand at Microsoft: Demand to Cancel
What we're witnessing is a very ungraceful destruction of XBox
 
Links 03/10/2025: Conflicts, Death of TypePad, and TikTok/CheeTok Gives a Boost to Far Right Groups in Europe
Links for the day
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Thursday, October 02, 2025
IRC logs for Thursday, October 02, 2025
Slopwatch: Linux Journal, Google News, and LinuxSecurity
They carry on polluting the Web with fake articles
Gemini Links 02/10/2025: Kubernetes With FreeBSD and robots.txt
Links for the day
Links 02/10/2025: 'Open' 'AI' Resorting to Gimmicks and Fake Funding, Europe’s ‘Drone Wall’ Discussed
Links for the day
Links 02/10/2025: Brave Passes 100M Users Milestone, Kodak Selling Its Own Film Again
Links for the day
Microsoft is Losing Europe
Hence all the "support" and "discount" offers that are limited to Europe
The Free Software Foundation Starts Fund-raising for 40th Anniversary
New pop-up 2-3 days ahead of the 40th anniversary event
Systemd Breaks Networking in Debian and Microsoft Staff Rushes to Make Face-Saving Excuses in LWN
Microsoft's bluca is already there in the comments, his Microsoft money pays for LWN to let him leave comments early
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Wednesday, October 01, 2025
IRC logs for Wednesday, October 01, 2025
What the End of XBox Will Look Like: a Fiery Crash
XBox is the next Skype. It won't last much longer. Expect many more layoffs.
Richard Stallman is Going to Finland to Give a Talk Next Thursday
A day later he speaks in Sweden
Gemini Links 02/10/2025: SMTP Pipelining and End of ROOPHLOCH 2025
Links for the day
Slopwatch: Plagiarism, Fake Articles, and FUD About Linux
not a day goes by without Google News feeding FUD from slopfarms
Gemini Links 01/10/2025: Chat Control and End of Life
Links for the day
Links 01/10/2025: Long Covid Risk Reiterated, "Bitcoin Queen" Caught
Links for the day
Links 01/10/2025: EA $55 Billion Deal is Debt and Slop "Raises Vishing Risks"
Links for the day
Bluewashing at Red Hat Means Redundancies
The man who sold Red Hat to IBM meanwhile became a Microsoft Mono booster
After Killing OpenSource.com, IBM ('Red Hat') and OSI Told Us OpenSource.net Would Replace It (But That Didn't Happen)
Now it's time to move on, perhaps tarnishing the "Open Source" label some more (for whatever sponsor wants this)
Linux is Not a Community Project, It's a Wall Street Product
The core goal should be freedom
Bad Actors Abusing the Free Software Community, Vandalising It Using Rogue Politics and Old Tactics
Oil giants have long attempted to do this; now, the digital equivalent of Big Oil does this in technology
Social Control Media Isn't the Future, The Federation or Fediverse Isn't Growing, People's Accounts Vanish for Good
users' accounts will get deleted, not just become inactive
IBM is Failing, This Helps Show Wall Street is Entirely Detached From Actual Commercial Performance
IBM is unable to grow, it's just constantly shrinking
Over at Tux Machines...
GNU/Linux news for the past day
IRC Proceedings: Tuesday, September 30, 2025
IRC logs for Tuesday, September 30, 2025
Clerical Aspects of Publishing and Development
In Free software, the management aspects are considerably reduced
Slopwatch: Fake Articles and Google News Promoting "Linux" Spam or Bot-Generated Fear, Uncertainty, Doubt (FUD)
These slopfarms help misplace blame
Third Wave of Microsoft Layoffs in September, This Time Many in Liverpool Affected
Be ready for more waves of layoffs ahead of the so-called "results" in late October