2
نموذجان للذكاء الاصطناعي
pre-trained tracker plus trained recogniser
أكاديمي · 2025
اكتب في الهواء بإصبعك، فيتحوّل ما تكتبه إلى نصّ رقمي. مساهمة في مشروع قائم ومفتوح المصدر في مجال الرؤية الحاسوبية، طوّرتُها وعرضتُها في الجامعة.
2
نموذجان للذكاء الاصطناعي
pre-trained tracker plus trained recogniser
على الجهاز نفسه
المعالجة
camera feed never leaves the machine
Academic
Context
contribution to an open-source project
المشكلة
Turning a fingertip traced through the air into recognised text joins two machine learning problems that do not naturally connect. Hand tracking produces a noisy time series of coordinates; handwriting recognition expects an image in the distribution it was trained on. The interesting work is the conversion between them.
الحلّ المعتمد
Took an existing open-source project and worked on top of it, extending the pipeline and presenting the result academically. The architecture combines a pre-trained hand landmark model with a handwriting recogniser trained on a standard academic dataset augmented with synthetic data, since off-the-shelf recognition is trained on scanned documents rather than thin single-stroke characters drawn in the air. The middle step is the hard one: a fingertip trajectory has to be smoothed against frame-to-frame jitter, resampled for frame-rate independence, and rendered and normalised into the canvas the recogniser expects.
بنية النظام
Two models in sequence: a pre-trained landmark detector for hand tracking, a purpose-trained recogniser for characters. Everything runs locally, so the camera feed never leaves the machine.
ومع كل قرار ثمنه
القرار الذي لا يُذكر ثمنه مجرّد تفضيل شخصي.
01
السبب
Hand landmark detection is solved and freely available. Recognising air-drawn strokes is not, because the input distribution is unlike any scanned handwriting dataset. Buy what is solved, build what is novel.
الثمن
A dependency on someone else model for the first stage, which is the correct trade against months spent reproducing it.
02
السبب
Real handwriting datasets are pen-on-tablet: steadier, with pen-up and pen-down signals air-writing does not have. Synthetic augmentation covers the gap the real data under-represents.
الثمن
Synthetic data can encode its own assumptions, which the validation split is there to expose.
النتيجة
A real-time pipeline from webcam to recognised text, with reported accuracy on both training and validation splits rather than a single headline number.