Detector de Falsificações

Assinaturas baseadas em imagens inseridas como sobreposições, DPI inconsistentes e campos de assinatura desalinhados.

Não — fornece sinais de risco heurísticos, não resultados forenses admissíveis em tribunal.

Não. Tudo é executado localmente no seu navegador.

Processe arquivos até o limite de memória do navegador.

Uma pontuação superior a 80 indica fortes sinais de manipulação de assinaturas baseadas em imagens.

The accuracy is high for exact or near-exact digital copies. OpenCV's normalized cross-correlation is a well-established computer vision technique used in industrial quality inspection and biometrics. False positives (legitimate signatures flagged as forgeries) can occur if two signers have virtually identical signatures or if the same person signed two documents identically. Always use this tool as a screening aid, not as definitive legal evidence.

Template matching is a computer vision technique that slides a template image across a larger image and measures how closely each region matches the template using a correlation score. OpenCV's matchTemplate function with the TM_CCOEFF_NORMED method produces a normalized score from -1 to +1, where +1 means a perfect match. The tool extracts the bottom 40% of one PDF page as the template and searches for it in the bottom 40% of the other — if the best match score exceeds the threshold, the signatures are structurally identical.