Air Signing and Privacy-Preserving Signature Verification for Digital Documents

TL;DR

This work proposes Air Signing, a camera-only, hardware-light approach to digital document signing that uses fingertip tracking to write signatures in mid-air and converts them into usable signature images. It combines Mediapipe-based fingertip localization with a Siamese network trained on the CEDAR and IN AIR SIGNATURE datasets to verify air signatures via a distance-based contrastive loss with margin $m$, producing a robust one-shot authentication framework. The study demonstrates competitive verification performance on the CEDAR dataset (accuracy ≈ $0.871$, FAR ≈ $5.39ackslash ext{%}$, FRR ≈ $7.48ackslash ext{%}$) and highlights weaker performance on in-air data, while showing a complete pipeline from real-time capture to signature storage. The approach enables privacy-preserving, hardware-free digital signing suitable for PDFs and online transactions, with potential extensions to natural-style signing and broader HCI applications.

Abstract

This paper presents a novel approach to the digital signing of electronic documents through the use of a camera-based interaction system, single-finger tracking for sign recognition, and multi commands executing hand gestures. The proposed solution, referred to as "Air Signature," involves writing the signature in front of the camera, rather than relying on traditional methods such as mouse drawing or physically signing on paper and showing it to a web camera. The goal is to develop a state-of-the-art method for detecting and tracking gestures and objects in real-time. The proposed methods include applying existing gesture recognition and object tracking systems, improving accuracy through smoothing and line drawing, and maintaining continuity during fast finger movements. An evaluation of the fingertip detection, sketching, and overall signing process is performed to assess the effectiveness of the proposed solution. The secondary objective of this research is to develop a model that can effectively recognize the unique signature of a user. This type of signature can be verified by neural cores that analyze the movement, speed, and stroke pixels of the signing in real time. The neural cores use machine learning algorithms to match air signatures to the individual's stored signatures, providing a secure and efficient method of verification. Our proposed System does not require sensors or any hardware other than the camera.

Figures (17)

• Figure 1: Datasets
• Figure 2: Table 2
• Figure 3: Hand Landmarks.
• Figure 4: Drawing postures (Active, Stop, Erase)
• Figure 5: Flow-Chart of Signing System