Decoding Process Research Articles

Mathematical model for 3D object reconstruction using OccNet (CNN)

The input 2D image is used by the encoder to first understand the geometrical restrictions in compressed representation. Second, in the straightforward Al method, the latent representation of the input image is acquired during encoding. On the other hand, the suggested OccNet (CNN) technique computes two encoded vectors of mean and standard deviation during the encoding stage from input. The acquired encoded representation is then transformed into a three-dimensional model via the decoding process. The same decoding process is used by both of the suggested solutions. The reconstruction of a complex 3D object with colourful effects from a single 2D shot may also be the subject of future research. Unlike other methods, our representation doesn’t need a lot of memory to encode a description of the 3D output at infinite resolution. We show that our representation effectively encodes three-dimensional structure and can be deduced from a variety of inputs. Our experiments show competitive results for the difficult challenges of 3D reconstruction from single images, noisy point clouds and coarse discrete voxel grids, both qualitatively and numerically.

Lw-Count: An Effective Lightweight Encoding-Decoding Crowd Counting Network

Crowd counting is a task of intelligent applications, and its operation efficiency is very important. However, in order to obtain a better counting performance, most of the existing works often design larger and more complex network structures, which will result in them occupying considerable memory, time and other resources at runtime, seriously limiting their deployment scope and making it difficult to be widely used in practical scenarios. In this paper, in order to overcome the above problems, we propose an effective lightweight encoding-decoding crowd counting network, named Lw-Count. Specifically, in the encoding process, we design an efficient and lightweight convolution module (ELCM), which extracts the crowd features in the network through a refined ghost block to reduce the network parameters and computing cost, and solves the problem of inaccurate counting caused by uneven crowd distribution through spatial group normalization (SGN). In the decoding process, we design a scale regression module (SRM) to reduce the error details and chessboard effect caused by linear interpolation and deconvolution. In addition, we design a new loss function, which enhances the spatial correlation of the density map and the sensitivity of the crowd through a regional normalized cross-correlation loss and counting loss, to ensure the counting accuracy. Extensive experiments on five mainstream datasets demonstrate that Lw-Count achieves an optimal trade-off between counting performance and running speed compared with other methods.

Collaborative Music Creation Method Based on Image Collection of Lattice Code

With the development of science and technology, more and more emerging technologies have been integrated into electronic music, forming a rich and diverse form of interactive electronic music. This research mainly discusses the collaborative music creation method based on bit matrix code image acquisition. In the decoding process, the displacement value of each point should be determined according to the direction and distance from the point to the reference point of the corresponding grid raster. This article marks the position characteristics of image symbols according to the dot matrix code notation defined in the improved coding method. After searching for a pair of dot codes with the shortest distance in the horizontal or vertical direction, starting from the first code point in the vertical direction and following the second code point. Through the above method, it is possible to accurately construct the grid raster lines of the dot matrix position encoding symbol image formed after the improved encoding, so that the decoding of the dot matrix symbol image is more accurate, and the composition translation is reduced. The improved coding method reduces the error of the origin matrix code by about 0.11 mm as a whole. This research will realize the collaborative recording of data in the music creation process, which will bring great convenience to creators.

Learned image compression with generalized octave convolution and cross-resolution parameter estimation

Recently, image compression approaches based on deep learning have gradually outperformed existing image compression standards including BPG and VVC intra coding. In particular, the application of the context-adaptive entropy model significantly improves the rate-distortion (R-D) performance, in which hyperpriors and autoregressive models are jointly utilized to effectively capture the spatial redundancy of the latent representations. However, the latent representations still contain some spatial correlations. In addition, these methods based on the context-adaptive entropy model cannot be accelerated in the decoding process by parallel computing devices, e.g. FPGA or GPU. To alleviate these limitations, we propose a learned multi-resolution image compression framework, which exploits the recently developed octave convolutions to factorize the latent representations into the high-resolution (HR) and low-resolution (LR) parts, similar to wavelet transform, which further improves the R-D performance. To speed up the decoding, our scheme does not use context-adaptive entropy model. Instead, we exploit an additional hyper layer including hyper encoder and hyper decoder to further remove the spatial redundancy of the latent representation. Moreover, the cross-resolution parameter estimation (CRPE) is introduced into the proposed framework to enhance the flow of information and further improve the rate-distortion performance. An additional information-fidelity loss is proposed to the total loss function to adjust the contribution of the LR part to the final bit stream. Experimental results show that our method separately reduces the decoding time by approximately 73.35 and 93.44 % compared with that of state-of-the-art learned image compression methods, and the R-D performance is still better than H.266/VVC(4:2:0) and some learning-based methods on both PSNR and MS-SSIM metrics across a wide bit rates.

Computational empowerment in practice: Scaffolding teenagers’ learning about emerging technologies and their ethical and societal impact

Whereas computational empowerment is well established in child–computer interaction research, less effort has been made to integrate this approach in practice. Building on four existing principles in child–computer interaction, namely Closeness, Embodied Learning, Design Process and Decoding, we explore how computational empowerment can be scaffolded in practice. This study was part of an interdisciplinary research project exploring emerging technologies in education (CEED). We involved 9 teachers and 26 students (14–15 y) in Denmark in teaching about machine learning and augmented reality. In the article we develop a framework for applying computational empowerment in formal educational contexts. Based on the study, we found that computational empowerment builds on distinct layers which are these four principles. Specifically, merging the principles Closeness and Embodied Learning can increase student engagement and motivation in the activities; in addition, Decoding and Design Process are key principles for promoting student awareness of the societal impacts of emerging technologies. Based on qualitative data from this intervention, we discuss the opportunities and challenges for child–computer interaction in moving toward computational empowerment in practice.

DRT-Unet: A Segmentation Network for Aiding Brain Tumor Diagnosis

Using image segmentation techniques to assist physicians in brain tumor diagnosis is a hot issue in computer technology research. Although most brain tumor segmentation networks to date have been based on U-Net, the prediction results are depending on which are not well generalized and need to be further improved. As the depth of the network increases, the gradients of the network vanish together with the decrease of the accuracy; meanwhile, the large number of parameters in the network will cause data redundancy. Moreover, a single modality of MRI images cannot adequately segment tumor details. Therefore, a segmentation network with an improved U-Net model is proposed in this paper, which combines Dilated Convolution-Dense Block-Transformation Convolution-Unet (hereafter referred to as DRT-Unet). The network adopts the combination of dilated convolution, dense residual block, and transposed convolution. In the coding process, a dilated convolution block and a local feature residual for fusing dense block are adopted to replace the 3 × 3 convolution layers on each layer in U-Net, and a transition layer is used for down-sampling. In the decoding process, a local feature residual is adopted for fusing dense blocks; meanwhile, a deconvolution structure with up-pooling and transposed convolution cascade is used. By connecting the decoded output features with the encoded low-level visual features, the information on transition layer loss is obtained. The experiments in this paper are carried out on BraTs2018 and BraTs2019 datasets; as a result, the DRT-Unet network can effectively segment tumor lesion regions.

MFI-Net: Multiscale Feature Interaction Network for Retinal Vessel Segmentation.

Segmentation of retinal vessels on fundus images plays a critical role in the diagnosis of micro-vascular and ophthalmological diseases. Although being extensively studied, this task remains challenging due to many factors including the highly variable vessel width and poor vessel-background contrast. In this paper, we propose a multiscale feature interaction network (MFI-Net) for retinal vessel segmentation, which is a U-shaped convolutional neural network equipped with the pyramid squeeze-and-excitation (PSE) module, coarse-to-fine (C2F) module, deep supervision, and feature fusion. We extend the SE operator to multiscale features, resulting in the PSE module, which uses the channel attention learned at multiple scales to enhance multiscale features and enables the network to handle the vessels with variable width. We further design the C2F module to generate and re-process the residual feature maps, aiming to preserve more vessel details during the decoding process. The proposed MFI-Net has been evaluated against several public models on the DRIVE, STARE, CHASE_DB1, and HRF datasets. Our results suggest that both PSE and C2F modules are effective in improving the accuracy of MFI-Net, and also indicate that our model has superior segmentation performance and generalization ability over existing models on four public datasets.

Reconfigurable Intelligent Surfaces: Simplified-Architecture Transmitters—From Theory to Implementations

Reconfigurable intelligent surfaces (RISs) offer an entirely new route to alter the propagation properties of electromagnetic waves and thus control their reflection, refraction, and scattering features in arbitrary manners. Such physical attributes are perceived to bring about fundamental influence on the modern wireless communication system due to the possibilities to establish artificial and controllable propagation environments for radio signals, which no longer rely on the complex encoding, decoding, and other signal processing techniques. Recent studies reveal that the wave manipulation is not the only skill of the RISs. With the rapid developments of space–time digital metasurface and information metasurface, there has been increasing attention focused on the information manipulation via these artificial surfaces. In this article, we provide an overview of the theoretical models of the space–time digital metasurface and information metasurface, the mechanisms of wavefront shaping, and the signal modulations in space and time domains during the wave–matter interactions. We will also address some practical issues during implementations of the reconfigurable intelligent metasurfaces and the associated hardware architectures at microwave frequencies to realize simplified radio frequency transmitters. Several modulation schemes and the corresponding demonstration systems are introduced to illustrate the powerful abilities of the reconfigurable intelligent metasurfaces. Potential research directions of this technique are briefly discussed for their potential applications in future wireless networks.

The Resignation of Wang An-Chi, the Artistic Director of Taiwan's GuoGuang Opera Company: A Debate from Two Versions of Phaedra (2019, 2021)

Taiwan's GuoGuang Jutuan (GuoGuang Opera Company, GGOC)'s artistic director, Wang An-Chi (b. 1955), resigned in November 2021 after serving in the position for almost twenty years. Two versions of GGOC's production of Phaedra in 2019 and 2021 led directly to her resignation, which opens up a series of questions about GGOC's creative path. This report firstly analyzes the choices made by GGOC in the two productions, and then explores the plight of GGOC as a state-owned and state-run jingju troupe in Taiwan which embodies the close relationship among jingju, Chineseness and national identity. In the event of Wang's resignation, I argue that the development of jingju in Taiwan is not a path to China (unification), but a series of continuous inquiries into Taiwan's subjective cultural identity. Jingju, as a case study of the transfer of the Taiwanese identity, showcases the unnecessarily coherent decoding and recoding process.

Performance Analysis and Optimization in Two-Way Cooperative NOMA Under Two-Phase Transmission Protocol

In this article, a novel two-phase transmission protocol is proposed to help the bidirectional transmission in a user-aid two-way cooperative nonorthogonal multiple access system, where superposition coding, bit-level Exclusive-Or operation, and successive interference cancellation (SIC) techniques are employed for improving the spectrum efficiency and throughput. Considering the residual interference generated in the practice process of decoding, both the perfect SIC and the imperfect SIC scenarios are discussed. Besides, the power allocation factor <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\alpha _{1}$</tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\alpha _{2}$</tex-math></inline-formula> are optimized under the perfect SIC to reduce the total outage probability, whose closed-form expression is derived and verified by simulations. Also, the approximated outage performance is derived in imperfect SIC scenario. Furthermore, two other conventional two-phase protocols and one three-phase protocol are provided as the benchmark to compare the outage probability and throughput with the proposed protocol, which shows the superiority of the proposed two-phase protocol and power allocation.

3D Reconstruction of Remote Sensing Mountain Areas with TSDF-Based Neural Networks

The remote sensing 3D reconstruction of mountain areas has a wide range of applications in surveying, visualization, and game modeling. Different from indoor objects, outdoor mountain reconstruction faces additional challenges, including illumination changes, diversity of textures, and highly irregular surface geometry. Traditional neural network-based methods that lack discriminative features struggle to handle the above challenges, and thus tend to generate incomplete and inaccurate reconstructions. Truncated signed distance function (TSDF) is a commonly used parameterized representation of 3D structures, which is naturally convenient for neural network computation and computer storage. In this paper, we propose a novel deep learning method with TSDF-based representations for robust 3D reconstruction from images containing mountain terrains. The proposed method takes in a set of images captured around an outdoor mountain and produces high-quality TSDF representations of the mountain areas. To address the aforementioned challenges, such as lighting variations and texture diversity, we propose a view fusion strategy based on reweighted mechanisms (VRM) to better integrate multi-view 2D features of the same voxel. A feature enhancement (FE) module is designed for providing better discriminative geometry prior in the feature decoding process. We also propose a spatial–temporal aggregation (STA) module to reduce the ambiguity between temporal features and improve the accuracy of the reconstruction surfaces. A synthetic dataset for reconstructing images containing mountain terrains is built. Our method outperforms the previous state-of-the-art TSDF-based and depth-based reconstruction methods in terms of both 2D and 3D metrics. Furthermore, we collect real-world multi-view terrain images from Google Map. Qualitative results demonstrate the good generalization ability of the proposed method.

Overview and Efficiency of Decoder-Side Depth Estimation in MPEG Immersive Video

This paper presents the overview and rationale behind the Decoder-Side Depth Estimation (DSDE) mode of the MPEG Immersive Video (MIV) standard, using the Geometry Absent profile, for efficient compression of immersive multiview video. A MIV bitstream generated by an encoder operating in the DSDE mode does not include depth maps. It only contains the information required to reconstruct them in the client or in the cloud: decoded views and metadata. The paper explains the technical details and techniques supported by this novel MIV DSDE mode. The description additionally includes the specification on Geometry Assistance Supplemental Enhancement Information which helps to reduce the complexity of depth estimation, when performed in the cloud or at the decoder side. The depth estimation in MIV is a non-normative part of the decoding process, therefore, any method can be used to compute the depth maps. This paper lists a set of requirements for depth estimation, induced by the specific characteristics of the DSDE. The depth estimation reference software, continuously and collaboratively developed with MIV to meet these requirements, is presented in this paper. Several original experimental results are presented. The efficiency of the DSDE is compared to two MIV profiles. The combined non-transmission of depth maps and efficient coding of textures enabled by the DSDE leads to efficient compression and rendering quality improvement compared to the usual encoder-side depth estimation. Moreover, results of the first evaluation of state-of-the-art multiview depth estimators in the DSDE context, including machine learning techniques, are presented.

Local-Partial Signal Combining Schemes for Cell-Free Large-Scale MU-MIMO Systems with Limited Fronthaul Capacity and Spatial Correlation Channels

Cell-free large-scale multi-user MIMO is a promising technology for the 5G-and-beyond mobile communication networks. Scalable signal processing is the key challenge in achieving the benefits of cell-free systems. This study examines a distributed approach for cell-free deployment with user-centric configuration and finite fronthaul capacity. Moreover, the impact of scaling the pilot length, the number of access points (APs), and the number of antennas per AP on the achievable average spectral efficiency are investigated. Using the dynamic cooperative clustering (DCC) technique and large-scale fading decoding process, we derive an approximation of the signal-to-interference-plus-noise ratio in the criteria of two local combining schemes: Local-Partial Regularized Zero Forcing (RZF) and Local Maximum Ratio (MR). The results indicate that distributed approaches in the cell-free system have the advantage of decreasing the fronthaul signaling and the computing complexity. The results also show that the Local-Partial RZF provides the highest average spectral efficiency among all the distributed combining schemes because the computational complexity of the Local-Partial RZF is independent of the UTs. Therefore, it does not grow as the number of user terminals (UTs) increases.

SC List-Flip Decoding of Polar Codes by Shifted Pruning: A General Approach.

In the successive cancellation (SC) list decoding, the tree pruning operation retains the L best paths with respect to a metric at every decoding step. However, the correct path might be among the L worst paths due to the imposed penalties. In this case, the correct path is pruned and the decoding process fails. shifted pruning (SP) scheme can recover the correct path by additional decoding attempts when decoding fails, in which the pruning window is shifted by paths over certain bit positions. A special case of the shifted pruning scheme where is known as SCL-flip decoding, which was independently proposed in 2019. In this work, a new metric that performs better in particular for medium and long codes is proposed, and nested shift-pruning schemes are suggested for improving the average complexity.

Lightweight human pose estimation algorithm based on polarized self-attention

In recent years, human pose estimation has been widely used in human-computer interaction, augmented reality, video surveillance, and many other fields, but the task of pose estimation still faces many challenges. To address the large number of parameters and complicated calculation in the current mainstream human pose estimation network, this paper proposes a lightweight pose estimation network (Lightweight Polarized Network, referred to as LPNet) based on a polarized self-attention mechanism. First, ghost convolution is used to reduce the number of parameters of the feature extraction network; second, by introducing the polarized self-attention module, the pixel-level regression task can be better solved, the lack of extracted features due to the decrease in the number of parameters can be reduced, and the accuracy of the regression of human keypoints can be improved; finally, a new coordinate decoding method is designed to reduce the error in the heatmap decoding process and improve the accuracy of keypoint regression. The method proposed in this paper was evaluated on the human keypoint detection datasets COCO and MPII, and compared with the current mainstream methods. The experimental results show that the proposed method greatly reduces the number of parameters of the model while ensuring a small loss in accuracy.

Cosinusoidal encoding multiplexed structured illumination multispectral ghost imaging.

The information dimension obtained by multispectral ghost imaging is more abundant than in single-band ghost imaging. Existing multispectral ghost imaging systems still meet some shortages, such as complex structure or reconstruction time-consuming. Here, an approach of cosinusoidal encoding multiplexed structured illumination multispectral ghost imaging is proposed. It can capture the multispectral image of the target object within one projection cycle with a single-pixel detector while maintaining high imaging efficiency and low time-consuming. The core of the proposed approach is the employed novel encoding strategy which is apt to decode and reconstruct the multispectral image via the Fourier transform. Specifically, cosinusoidal encoding matrices with specific frequency characteristics are fused with the orthogonal Hadamard basis patterns to form the multiplexed structured illumination patterns. A broadband photomultiplier is employed to collect the backscattered signals of the target object interacted by the corresponding structured illumination. The conventional linear algorithm is applied first to recover the mixed grayscale image of the imaging scene. Given the specific frequency distribution of the constructed cosinusoidal encoding matrices, the mixed grayscale image can be converted to the frequency domain for further decoding processing. Then, the pictures of multiple spectral components can be obtained with some manipulations by applying Fourier transform. A series of numerical simulations and experiments verified our proposed approach. The present cosinusoidal encoding multiplexed structured illumination can also be introduced in many other fields of high-dimensional information acquisition, such as high-resolution imaging and polarization ghost imaging.

Nondestructive Sequencing of Enantiopure Oligoesters by Nuclear Magnetic Resonance Spectroscopy.

Sequence-defined synthetic oligomers and polymers are promising molecular media for permanently storing digital information. However, the information decoding process relies on degradative sequencing methods such as mass spectrometry, which consumes the information-storing polymers upon decoding. Here, we demonstrate the nondestructive decoding of sequence-defined oligomers of enantiopure α-hydroxy acids, oligo(l-mandelic-co-d-phenyl lactic acid)s (oMPs), and oligo(l-lactic-co-glycolic acid)s (oLGs) by 13C nuclear magnetic resonance spectroscopy. We were able to nondestructively decode a bitmap image (192 bits) encoded using a library of 12 equimolar mixtures of an 8-bit-storing oMP and oLG, synthesized through semiautomated flow chemistry in less than 1% of the reaction time required for the repetition of conventional batch reactions. Our results highlight the potential of bundles of sequence-defined oligomers as efficient media for encoding and decoding large-scale information based on the automation of their synthesis and nondestructive sequencing processes.

Chunking in Visual Working Memory: Are Visual Features of Real-World Objects Stored in Chunks?

Are visual features of real-world objects stored as bound units? Previous research has shown that simple visual features (e.g., colored squares or geometric shapes) can be effectively bound together when forming predictable pairs in memory tasks. Through a "memory compression" process, observers can take advantage of these features to compress them into a chunk. However, a recent study found that visual features in real-world objects are stored independently. In the present study, we explored this issue by using drawings of fruits as memory stimuli, presenting four pictures of fruit in separate test trials in which we required observers to remember eight total features (i.e., four colors and four shapes). In the congruent trials, the color of the fruit matched its natural appearance (e.g., a red apple), while in incongruent trials, the color of the fruit mismatched its natural appearance (e.g., a red banana). We paired the shape of the fruits randomly with a color (without replacement). According to chunking theory, if visual features of real-world objects are stored in a chunk, the highest memory capacity should be accompanied by the longest response time in congruent trials due to an extra decoding process required from the chunk. We did find that participants had the highest memory capacity in the congruent condition, but their response times in the congruent condition were significantly faster than in the incongruent condition. Thus, observers did not undergo a decoding process in the congruent condition, and we concluded that visual features in real-world objects are not stored in a chunk.

Robust OCC System Optimized for Low-Frame-Rate Receivers

Light emitting diodes (LED) are becoming the dominant lighting elements due to their efficiency. Optical camera communications (OCC), the branch of visible light communications (VLC) that uses video cameras as receivers, is a suitable candidate in facilitating the development of new communication solutions for the broader public because video cameras are available on almost any smartphone nowadays. Unfortunately, most OCC systems that have been proposed until now require either expensive and specialized high-frame-rate cameras as receivers, which are unavailable on smartphones, or they rely on the rolling shutter effect, being sensitive to camera movement and pointing direction, they produce light flicker when low-frame-rate cameras are used, or they must discern between more than two light intensity values, affecting the robustness of the decoding process. This paper presents in detail the design of an OCC system that overcomes these limitations, being designed for receivers capturing 120 frames per second and being easily adaptable for any other frame rate. The system does not rely on the rolling shutter effect, thus making it insensitive to camera movement during frame acquisition and less demanding about camera resolution. It can work with reflected light, requiring neither a direct line of sight to the light source nor high resolution image sensors. The proposed communication is invariant to the moment when the transmitter and the receiver are started as the communication is self-synchronized, without any other exchange of information between the transmitter and the receiver, without producing light flicker, and requires only two levels of brightness to be detected (light on and light off). The proposed system overcomes the challenge of not producing light flicker even when it is adapted to work with very low-frame-rate receivers. This paper presents the statistical analysis of the communication performance and discusses its implementation in an indoor localization system.

Spatial–temporal phase demodulation decoding superimposed ghost reflections in optical testing

We introduce a method of phase demodulation to distinguish the ghost reflection commonly observed in optical testing. The method digitally processes a sequence of fringe patterns to separate the reflections from the front and rear surfaces. First, we use a phase-shifting algorithm to compute an analytic signal with the two reflections coupled, and then we decouple them through a maximized spatial filtering. The process requires the acquisition of at least three fringe patterns, as well as fulfilling opposite signs for the magnifications of the two reflections. We demonstrate the decoding process through a real phase measuring deflectometry data processing example.