Restoring 2D Content from Distorted Documents

Abstract

Abstract-This paper presents a framework to restore the 2D content printed on documents in the presence of geometric distortion and nonuniform illumination. Compared with text-based document imaging approaches that correct distortion to a level necessary to obtain sufficiently readable text or to facilitate optical character recognition (OCR), our work targets nontextual documents where the original printed content is desired. To achieve this goal, our framework acquires a 3D scan of the document's surface together with a high-resolution image. Conformal mapping is used to rectify geometric distortion by mapping the 3D surface back to a plane while minimizing angular distortion. This conformal "deskewing" assumes no parametric model of the document's surface and is suitable for arbitrary distortions. Illumination correction is performed by using the 3D shape to distinguish content gradient edges from illumination gradient edges in the high-resolution image. Integration is performed using only the content edges to obtain a reflectance image with significantly less illumination artifacts. This approach makes no assumptions about light sources and their positions. The results from the geometric and photometric correction are combined to produce the final output.

Document dewarping via text-line based optimization

“…Since the 3D structure of an object can be estimated by using structured lights or laser scanners, many document dewarping methods were developed based on shape measuring hardwares [8,9,20,21]. Although these approaches allow us to handle a range of surface distortions and yield accurate results, the need of special equipments is their major drawback.…”

“…From the ground-truth distortion parameters, we first compute the true dewarping functions in (8), and compare the function with the estimated one. The error metric is given by…”

“…However, the camera-captured document images usually suffer from geometric distortions caused by capturing processes (perspective distortions) and curved surfaces. In order to reduce these geometric distortions, a number of approaches have been proposed [5][6][7][8][9][10][11][12][13][14][15][16]. Among various approaches, the text-line based approach has received a lot of attention, because text-lines are the most salient properties of documents and this approach allows us to develop an algorithm that needs a single image as the input.…”

Asset digitization

“…Its vellum pages are cockled and therefore show distortion under attempts at standard photography. The digitally flattened result is acquired from a digital photograph that is texture mapped onto an exact shape model of the page after running the shape model through a material simulation to semi-rigidly flatten it [Brown et al 2007]. The resulting image reduces or removes the distortions caused by shape and shading artifacts in the original photograph.…”

Retrospective Illumination Correction of Retinal Images

“…Other approaches exist, for example, in applications including illumination correction for face recognition [23, 24] and restoration of digitized documents [25, 26]. The latter application frequently uses multiplicative model and specific properties of the processed images (e.g., illumination edges or geometrical distortion).…”