A Recursive Edge Detector For Color Filter Array Image

Abstract

Most of embedded cameras use a single sensor to capture images through a color filter. They produce special images with only one color component per pixel. Missing data are usually estimated through a demosaicking process, but this takes undesirable computation time and may generate undesirable color artifacts. Many embedded systems under real-time constraints could use this type of camera for purposes like edge detection. In this paper, a new edge detection method is proposed for the computation of partial derivative images. This algorithm is tested on a large set of synthetic images where edge ground truth is unquestionable. The exploitation of the raw data of images allows not only to drastically reduce computational time, but also to get edge detection results even more precise than using certain demosaicking-based methods.

Forgered Image Perception System using CNN Algorithms

“…Mahwatta et al describe the implementation steps, experimental results, and discusses the effectiveness of the proposed method in detecting image splicing. Magnier et al (2020) [5] introduces a novel approach to edge detection directly from Color Filter Array (CFA) images, bypassing the demosaicking step. The paper reviews traditional demosaicking-based edge detection methods and proposes a method leveraging recursive Deriche filters for computing partial derivative images.…”

Shen-Castan Based Edge Detection Methods for Bayer CFA Images

“…In the works of Aberkane et al [4] and Magnier et al [5], Deriche-based algorithms were proposed (Eqs. ( 5) and ( 6)) directly on the CFA image to extract contours.…”

“…On another note, the DRGC (Deriche derivative of Rotated Green-Chrominance) method proposed by Magnier et al [5] operates on Bayer CFA images in the diagonal directions in order to obtain pixel continuity (see Fig. 1).…”

“…2.2) involves the demosaicing process, corresponding to 30 to 1212 steps per pixel depending on the algorithm [3][4]. Note also that the DRGC method has 2×3 rotation steps [5]. Thus, as described in [13], the normalized N edge detection evaluation measure is, for F N >0 or F P >0:…”

“…Consequently, for simplicity and convenience, features may be extracted using grayscale vision sensors indicating only the brightness of a pixel. In response to the demand of feature extraction efficiency directly on CFA images, Aberkane et al [4] and Magnier et al [5] have proposed edge detection algorithms that operate based on Deriche edge detector and bilinear interpolations [6].…”