Since 2014 we have been building algorithms that normalise, frame and verify portrait photographs for passports, identity cards and badges. Available as SDKs to embed in your own software, or as ready-to-use applications for enrolment desks.
The ICAO compliance check really runs: upload one of your own photographs and our SDK analyses it in real time. For the other products you will find interactive walk-throughs built from real output of the desktop application — the very same screens you would see on your monitor.
ICAO compliance check — real processing
Upload a photo from your device or take one with the camera: the image is analysed by our SDK and you get the verdict on each individual ICAO requirement, together with guidance on how to fix the ones that fail. It also works from a smartphone.
Server unreachable The processing service is not answering right now: the button opens the interactive walk-through built from real, previously computed results.
Face detection, landmarks, compliant crop
Five real cases, from an already framed headshot to a full-length photograph with a tilted head: you can see the facial landmarks, the inter-pupillary distance, the rotation correction and the final compliant crop.
Enrolment desk, before/after table
Three real subjects with the original / primary / black-and-white secondary triptych, plus the table of ICAO checks before and after processing: you can read exactly which requirements were out of specification and how they came back in.
Normalisation pipeline via SDK
The normalisation sequence applied to a real sample, with the actual processing times measured on the desktop application.
Three different subjects, each with the table of ICAO requirements measured before and after normalisation: you can read row by row what was out of specification and how it came back in.
The same algorithms, packaged differently depending on where they are needed: inside your own application, on an enrolment workstation, or across an existing archive of images to be reprocessed in bulk.
Compliance verification SDK
Analyses a photograph and reports the outcome for every ICAO 9303 and ISO/IEC 19794-5 requirement: face geometry, pose, gaze, lighting, background, sharpness and colour rendition. It embeds KeeSquare technology, used in national document issuance programmes.
Compliance checking, second generation
The newest application from Digit-Ale.com. PNICAO (v2) is currently being trained on 500,000 ICAO passport-photo images, and will also cover the ISO 39794-5 standard.
W.I.P. Under development: no demo and no datasheet yet.
First stage: find and frame
No pre-cropped photograph needed. From a close-up or a full-length shot it locates the face, extracts the landmarks (5 or 68 points), measures the inter-pupillary distance, straightens a tilted head and produces a compliant crop in the required format.
Application for enrolment desks
A desktop interface for staff capturing photographs at the counter, connectable to scanners, cameras, webcams or network folders. Customisable and designed for operators without technical training: load the image, get the compliant photo.
Pure normalisation, batch capable
The normalisation and equalisation stage alone, for images that are already centred and cropped. It connects to existing databases to reprocess whole photographic archives in bulk. No training required.
W.I.P. Interactive walk-through on hold: the screenshots currently available do not represent the final product interface. Awaiting the correct material.
Integration SDK
The development kit for grafting normalisation into an application that already exists, printing solutions included. For languages other than .NET we supply the pure normalisation algorithms, which can be integrated on any platform.
Work archive
A collection of the projects delivered from 2014 onwards, with the results obtained in the field.
W.I.P. Section to be built: past projects will live here, in a dedicated archive, rather than scattered across the product pages.
The products are stages of the same chain. Use them all in sequence, or only the one your process is missing.
Face detection, facial landmarks, inter-pupillary distance measurement, tilt correction and cropping to the required format.
PhotoNorm Face FinderColour and brightness equalisation, skin tone rendition, background treatment: the algorithmic stage that makes even a badly taken photograph usable.
PhotoNorm Base / EnrollICAO 9303 and ISO/IEC 19794-5 requirements checked one by one, with the outcome and an indication of what is wrong.
Digit-Ale MorpheusICAOColour primary image and black-and-white secondary image, in Passport and ID card formats, ready for printing or archiving.
OutputThe sample image sets contain both “ordinary” portraits and difficult cases that are unusable without normalisation: they are meant to stress the algorithms, not to flatter them.
The Windows demonstration applications are being updated: the previous builds have been withdrawn from download because they no longer represent the current products. To receive them, together with a 90-day trial licence for a single workstation, write to info@digit-ale.com.
Digit-Ale.com develops and markets the PhotoNorm and Digit-Ale MorpheusICAO suites, available as applications, modules or SDKs. The core of our work is the normalisation of portrait photographs intended for identity documents: making any photograph usable, with or without operator intervention, including inside fully automated processes.
This section is to be rewritten with the up-to-date history of the business and its references, which will live in a dedicated project archive. Awaiting final content.
For commercial enquiries, technical evaluations, or to receive the demonstration applications with a trial licence.
