Convolutional code | EPFL Graph Search

In telecommunication, a convolutional code is a type of error-correcting code that generates parity symbols via the sliding application of a boolean polynomial function to a data stream. The sliding application represents the 'convolution' of the encoder over the data, which gives rise to the term 'convolutional coding'. The sliding nature of the convolutional codes facilitates trellis decoding using a time-invariant trellis. Time invariant trellis decoding allows convolutional codes to be maximum-likelihood soft-decision decoded with reasonable complexity. The ability to perform economical maximum likelihood soft decision decoding is one of the major benefits of convolutional codes. This is in contrast to classic block codes, which are generally represented by a time-variant trellis and therefore are typically hard-decision decoded. Convolutional codes are often characterized by the base code rate and the depth (or memory) of the encoder [n,k,K]. The base code rate

Efficient error correction solutions for OFDM based digital video

Mauro Lattuada

The robustness of the delivery of digital compressed video streams has always been one of the main topic of the standardization bodies to assure the Quality of the Service (QoS). The MPEG standard which is nowadays the reference for digital television defines how to format the various component parts of a multimedia program and how they are combined into a single transmission bit stream. Although the MPEG stream may be directly used over a wide variety of media, it may also be used over a communication network and it is already packetized into short data packets. The multiplexed stream can be transmitted over a variety of communication networks such as Radio Frequency links (UHF/VHF), Digital Satellite links, Cable TV Networks, Standard Terrestrial Communication Links (PDH, SDH) or Packet/Cell Links (ATM, IP, IPv6, Ethernet). In critical and error prone channels the robustness of the stream has to be granted by employing strong error correction schemes. Depending on the used media the transmission has its own coding scheme and modulation methods designed for the particular environment. For example, in the case of IP based links the transfer protocol accepts loss of data and the problem can be easily overcome by retransmission. Wireless simplex transmission cannot exploit the same approach. Wireless transmission always means variable channel conditions especially when one or both terminals are moving on the outfield. The channel fading evolution and the signal impairments increase the error rate and corrupt the data. In the recent past the standardization of the DVB-T shows the possibility to provide a robust service for digital video delivery. This standard provides superior picture quality with the opportunity to view pictures in standard format or wide screen (4/3 or 16/9) formats, services including subtitling, multiple audio tracks, interactive content, multimedia content where, for instance, programs may be linked to world wide web material. Beside the provided services and features the standard offers also a modulation and transmission scheme capable to cope with the impairments of the terrestrial channel and to provide a BER after forward-error correction below 10-12. Unfortunately, these performances are guaranteed only for fixed reception, since the standard was originally aimed to replace classical analog television broadcasting in the UHF band. This thesis addresses the study of enhanced error-correction schemes in the scope of a compact, flexible and robust high-throughput real time video transmission system able to cope with high mobility channels. The study of a transmission system and error correction scheme tailored for this kind of application is here presented. The system exploits the OFDM modulation technique. This modulation scheme has already proven to be very effective in hostile environments. A flexible modulation scheme is employed to obtain variable bandwidth transmission and enhanced performances especially for mobile applications. The main topic of this research is the study of an effective error correction scheme adapted to this application. While nowadays the digital television standards use a concatenated error correction scheme based on convolutional inner coder, in this work the proposed solution is based on a turbo inner coder. Turbo codes has already shown good performances compared to convolutional codes. These results are here applied to a concatenated scheme to obtain very low BER. A comparison of the performances of various turbo coding techniques under certain implementation constraints is done in this work.The second element of the error correction scheme is the channel interleaver used to average the channel fading in time and frequency. This mechanism is conceived to be adapted to the modulated bandwidth and is optimized for high throughput real time video. This study shows that the proposed system proves to be very effective in high mobile channels because of the deep averaging strategies in conjunction with the high efficiency turbo based concatenated error correction scheme.

EPFL2005

DermoNet: densely linked convolutional neural network for efficient skin lesion segmentation

Saleh Baghersalimi, Hazim Kemal Ekenel, Jean-Philippe Thiran

Recent state-of-the-art methods for skin lesion segmentation are based on convolutional neural networks (CNNs). Even though these CNN-based segmentation approaches are accurate, they are computationally expensive. In this paper, we address this problem and propose an efficient fully convolutional neural network, named DermoNet. In DermoNet, due to our densely connected convolutional blocks and skip connections, network layers can reuse information from their preceding layers and ensure high accuracy in later network layers. By doing so, we take advantage of the capability of high-level feature representations learned at intermediate layers with varying scales and resolutions for lesion segmentation. Quantitative evaluation is conducted on three well-established public benchmark datasets: the ISBI 2016, ISBI 2017, and the PH2 datasets. The experimental results show that our proposed approach outperforms the state-of-the-art algorithms on these three datasets. We also compared the runtime performance of DermoNet with two other related architectures, which are fully convolutional networks and U-Net. The proposed approach is found to be faster and suitable for practical applications.

2019

DermoNet: Densely Linked Convolutional Neural Network for Efficient Skin Lesion Segmentation

Seyedbehzad Bozorgtabar, Jean-Philippe Thiran

Recent state-of-the-art methods for skin lesion segmentation are based on Convolutional Neural Networks (CNNs). Even though these CNN based segmentation approaches are accurate, they are computationally expensive. In this paper, we address this problem and propose an efficient fully Convolutional Neural Network, named DermoNet. In DermoNet, due to our densely connected convolutional blocks and skip connections, network layers can reuse information from their preceding layers and ensure high accuracy in later network layers. By doing so, we take advantage of the capability of high-level feature representations learned at intermediate layers with varying scales and resolutions for lesion segmentation. Quantitative evaluation is conducted on three well-established public benchmark datasets: the ISBI 2016, ISBI 2017, and the PH2 datasets. The experimental results show that our proposed approach outperforms state-of-the-art algorithms on these three datasets. We also compared the runtime performance of DermoNet with two other related architectures, which are FCN and U-Net. The proposed approach is found to be faster and suitable for practical applications.

2018

Efficient error correction solutions for OFDM based digital video

Mauro Lattuada

EPFL2005