Optimization Design of Polar-LDPC Concatenated Scheme Based on Deep Learning

The concatenation of polar codes with LDPC codes is a promising approach to enhance the performance of polar codes at finite lengths while suppressing the error floor of LDPC codes. The polarized bit-channels of polar coding at finite lengths exhibit different qualities, i.e., some are almost noiseless while others are noisy. This implies that unbalanced protection is needed in the design of concatenating polar codes with other codes. To this end, a scheme of concatenatng polar codes with outer irregular LDPC codes for the protection of intermediate channels of polar codes is proposed in this paper. Between two constituent codes, a specifically designed interleaver is employed such that the overall error probability is minimized. The extrinsic information transfer (EXIT) analysis and a low-complexity heuristic method are proposed for the interleaver design. The soft cancellation (SCAN) and belief propagation(BP) algorithms are employed for the decoding of concatenated codes. Simulations are provided to validate its advantages in terms of excellent error correction capability and low decoding complexity.

Interleaver Design for LDPC-Partial Polar Codes Based on EXIT Analysis

Noise interference during the acquisition of digital images can severely degrade image quality, particularly for images captured under low-light conditions; however, the removal of image noise requires sophisticated digital image processing systems. This study presents a hardware-based solution to real-time image denoising using an existing algorithm designed for the removal of mixed impulse noise (salt-and-pepper and random-valued impulse noise), while preserving image edge details and image borders, without the need for additional computation time or memory capacity. Note that mixed impulse noise is typical of most real-world situations, such as the video noise associated with dashboard cameras. The proposed design was implemented using 180 nm complementary metal-oxide-semiconductor (CMOS) technology, consuming only 21.7 mW when operated at 200 MHz. This operating frequency allows the proposed chip to process noisy video streams with resolution of $1920\mathbf {\times }1080$ at 60 frames per second in real time. In terms of image restoration, the proposed algorithm achieved image quality on par with that achieved using software simulation. We also demonstrated the efficacy of the proposed scheme in denoising noisy video images from a dashboard camera.

Realization of a Real-Time Image Denoising System for Dashboard Camera Applications

We present an adaptive threshold method for salt and pepper noisy fingerprint images. At first, we applied a threshold algorithm to detect noisy and corrupt pixels. Fingerprint image optimized with window filter produced pixels values. We measured the quality of the fingerprint images with peak signal to noise ratio and other efficient noise removing techniques were compared with it. The result shows improving performance of our method with remove salt and pepper noise from a fingerprint image. Our present work is useful to reduce or minimize medium level noise in fingerprint image. In future, we will work to reduce more than 80 percent noise level in fingerprint image.

https://iopscience.iop.org/article/10.1088/1757-899X/1077/1/012049/pdf

Fingerprint Salt & Pepper Noisy Image Enhancement with Threshold Method

In this paper we consider the special construction of concatenated codes. Our scheme consists of outer LDPC code and a specially designed class of inner LDPC code with interleaver between inner and outer codes. An inner code in the proposed construction is obtained using a specific optimization algorithm that allows to minimize SNR for a given target frame error probability. The decoding of the proposed codes is based on the idea of woven codes decoding. Simulation results show that suggested code construction being decoded by woven-based decoding algorithm outperforms low-rate 5G LDPC codes. 

On the woven-like concatenation of LDPC codes

This paper reports low bit-error-rate (BER) polar codes concatenated low density parity check (LDPC) codes for wireless communication systems. The LDPC code has been validated on the error correction capable to approaching the theoretical value. However, the BER performance of this code is still limited because of the effect of the error floor. Therefore, a concatenated coding scheme based on polar and LDPC codes has been proposed to further reduce the BER. Compared to Reed-Solomon codes concatenated with LDPC codes, this polar-LDPC concatenated coding has better BER performance.

Lower bit-error-rate polar-LDPC concatenated coding for wireless communication systems

In the transmission, digital images often disturbed by various interferences, leading to the image corrupted severely. For the better visual quality and the further use in digital image processing systems, noises must be removed. In this study, the proposed scheme can efficiently remove fixed-valued and random-valued impulse noise sources while preserving image edges. It has better image quality than those of the conventional linear and non-linear filters. The proposed scheme also maintains the edge of the restoration image due to preserving image edge details. Because its procedure is simple, it is suited for real-time applications and hardware implementation.

Zhi-Hong Lin

Papers

Lower bit-error-rate polar-LDPC concatenated coding for wireless communication systems

Efficient image denoising scheme for removal of impulse noise