Color Seal Extraction from Documents: Robustness through Soft Data Fusion

Abstract

This paper presents a framework for the extraction of elements characterized by a particular color hue from color document images. The presented approach attains the detection of official seals, which will be thence analyzed by an embedding system in order to detect possible falsifications. The framework is based on the fusion operator denoted as fuzzy integral, whose robustness with respect to changes in the luminance and the saturation of particular hues is due to the use of the ranking among the input data as an influencing factor in the fusion result. The approach is evaluated on a real data set of tax forms delivered by custom houses, showing its successful performance

Fuzzy fusion for skin detection

“…1 the fuzzy integral is used for skin detection both within the DoFI methodology [15], and for the fusion of three estimates of the skin distribution on the input images. These two approaches based on the employment of the fuzzy integral are described in the following two sections.…”

“…As formerly mentioned a methodology based on the fuzzy integral was used for the segmentation of ink seals in [15] and denoted as DoFI. The main feature of this methodology is its ability to isolate a cluster characterized by a particular color hue from an input image.…”

“…This property results in the robustness of the operator w.r.t. a change in the luminance/saturation of color hues, which has been exploited in [15] for the detection of ink seals on document images. Taking all these facts into account a framework for the robust skin detection that is based on the application of the fuzzy integral is presented herein.…”

“…First the methodology for the detection of particular color clusters presented in [15] is applied. This methodology, which was denoted as Difference of Fuzzy Integrals (DoFI), is applied in order to estimate the skin distribution in the input images.…”

A New Seal Verification For Chinese Color Seal

“…Several seal verification methods have also been proposed for decades [3][4]. These methods mainly include four steps: 1) Rotate the candidate seal imprint to the same orientation as the reference one; 2) Compute the centroid to the candidate imprint and reference one; 3) Overlap the rotated candidate seal imprint and the reference one in terms of the centroid; 4) Match them pixel by pixel.…”