Compressed Blocks Research Articles

ICC-BiFormer: A Deep-Learning Model for Near-Earth Asteroid Detection via Image Compression and Local Feature Extraction

Detecting near-Earth asteroids (NEAs) is crucial for research in solar system and planetary science. In recent year, deep-learning methods have almost dominated the task. Since NEAs represent only one-thousandth of the pixels in images, we proposed an ICC-BiFormer model that includes an image compression and contrast enhancement block and a BiFormer model to capture local features in input images, which is different from previous models based on Convolutional Neural Network (CNN). Furthermore, we utilize a larger input size of the model, which corresponds to the side length of the input image matrix, and design a cropping algorithm to prevent NEAs from being truncated and better divide NEAs and satellites. We apply our ICC-BiFormer model into a dataset of approximately 20,000 streak and 40,000 non-streak images to train a binary classification model. The ICC-BiFormer achieves 99.88% accuracy, which is superior to existing models. Focusing on local features has been proven effective in detecting NEAs.

ScCompressSA: dual-channel self-attention based deep autoencoder model for single-cell clustering by compressing gene-gene interactions.

Single-cell clustering has played an important role in exploring the molecular mechanisms about cell differentiation and human diseases. Due to highly-stochastic transcriptomics data, accurate detection of cell types is still challenged, especially for RNA-sequencing data from human beings. In this case, deep neural networks have been increasingly employed to mine cell type specific patterns and haveoutperformed statistic approachesin cell clustering. Using cross-correlation to capture gene-gene interactions, this study proposes the scCompressSA method to integrate topological patterns from scRNA-seq data, with support of self-attention (SA) based coefficient compression (CC) block. This SA-based CC block is able to extract and employ static gene-gene interactions from scRNA-seq data. This proposed scCompressSA method has enhanced clustering accuracy in multiple benchmark scRNA-seq datasets by integrating topological and temporalfeatures. Static gene-gene interactions have been extracted as temporal features to boost clustering performance insingle-cell clustering For the scCompressSA method, dual-channel SA based CC block is able to integrate topological features and has exhibited extraordinary detection accuracy compared with previous clustering approaches that onlyemploy temporal patterns.

Compressed earthen blocks using alluvial clays from Mbam: performance comparison using statistical analysis of cement vs heat-stabilized blocks

AbstractIn the present study, a comparison of the thermal-insulation and mechanical performances of cement and heat-stabilized compressed earthen blocks (CEBs) was carried out to determine the factors which influence those properties. The raw clays used consist mainly of kaolinite, orthoclase and quartz. The mechanical strength increased with increase in both the amount of cement added and the firing temperature. However, the responses are better for cement-stabilized CEBs. The thermal insulation of fired bricks is greater than that of cement-stabilized bricks. This difference was related to the decrease in porosity and the formation of continuous-surface. The decrease in thermal insulation is mainly related to the formation of continuous-surface in cement-stabilized CEBs, whereas in the fired CEBs, it is due to the modification of pore volume. The mineralogy of the raw clays is statistically correlated to porosity and continuous-surface development that were confirmed as the main factors in the modification of both the mechanical strength and the thermal insulation. In cement-stabilization, the decrease in insulation is due to the development of continuous surface, while for heat-stabilization, mineral transformations during the sintering reduced continuous-surface formation and the insulation was controlled by both radiation and reduced surface conduction. The influence of the mineralogy of the raw material shows that clay content favours the insulation in fired bricks obtained at T ≤ 1000°C, while sand contents favour densification. In contrast, clay contents reduce the mechanical response of cement-stabilized material due to limited cement–clay interactions. In general, the mechanical response is more favourable in cement stabilization, while thermal insulation is better in fired bricks.

Real‐Time Neural Materials using Block‐Compressed Features

AbstractNeural materials typically consist of a collection of neural features along with a decoder network. The main challenge in integrating such models in real‐time rendering pipelines lies in the large size required to store their features in GPU memory and the complexity of evaluating the network efficiently. We present a neural material model whose features and decoder are specifically designed to be used in real‐time rendering pipelines. Our framework leverages hardware‐based block compression (BC) texture formats to store the learned features and trains the model to output the material information continuously in space and scale. To achieve this, we organize the features in a block‐based manner and emulate BC6 decompression during training, making it possible to export them as regular BC6 textures. This structure allows us to use high resolution features while maintaining a low memory footprint. Consequently, this enhances our model's overall capability, enabling the use of a lightweight and simple decoder architecture that can be evaluated directly in a shader. Furthermore, since the learned features can be decoded continuously, it allows for random uv sampling and smooth transition between scales without needing any subsequent filtering. As a result, our neural material has a small memory footprint, can be decoded extremely fast adding a minimal computational overhead to the rendering pipeline.

Dynamic Response of Pile-Raft Systems with Various Forms of Connection under Cyclic Condition

This study aimed to examine the aseismic performance of the pile-raft systems with various connection forms. The related shaking table test and numerical simulation were performed for in-depth investigation. The acceleration response spectra on the top of the soil layer and raft were obtained and plotted for contrastive analysis based on the model test at a reduced scale of 1 : 30 and finite element numerical simulation. Accordingly, the working mechanisms of the pile-raft systems in conventional connections with the embedment of the compressible blocks and cushion layers under cyclic loading were explored. The results showed that the embedment of the cushion layer on the pile top could most significantly mobilize the potential of the foundation soil, effectively reduce the bending moment peak of the pile, and reduce the acceleration amplification effect on the top of the soil layer and raft. The embedment of the compressible block on the pile top most markedly reduced the bending moment peak of the pile and effectively mobilized the potential of the foundation soil, which was most favorable for lowering the amplification effect of acceleration on the top of the soil layer and raft.

Finite Element Simulation of Compressing an Additively Manufactured Mesostructure

Abstract Numerical modeling is a useful approach for answering research questions that are either difficult or inefficient to answer experimentally. The goal of this study was to develop a validated numerical model of a stereolithography (SLA) manufactured mesostructure to predict compression behavior. A three-dimensional (3D) model was created for an 80 × 80 × 20 mm compression block. ANSYS was used to simulate compression testing of this structure. The lower plate of the structure was fixed while a vertical displacement was prescribed on the upper plate. The base material was modeled using a linearly elastic, isotropic material derived from Young's modulus (E), Poisson's ratio, and the ultimate tensile strength. The base material was varied to represent three different SLA polymers from Formlabs (i.e., three different models of the same geometrical structure). The shape of the force–displacement curves and compression force values agreed well between the model and previously collected experimental data. However, this agreement was limited to stiffer materials and only up to 10 mm of compression. This may be where the structure shifts from elastic to plastic behavior and some lattice structures fracture. This simulation tool can be used in the future to predict and optimize the behavior of this lattice structure but only while operating elastically.

Ultra-low latency spiking neural networks with spatio-temporal compression and synaptic convolutional block

Spiking neural networks (SNNs), as one of the brain-inspired models, has spatio-temporal information processing capability, low power feature, and high biological plausibility. The effective spatio-temporal feature makes it suitable for event streams classification. However, neuromorphic datasets, such as N-MNIST, CIFAR10-DVS, DVS128-gesture, need to aggregate individual events into frames with a new higher temporal resolution for event stream classification, which causes high training and inference latency. In this work, we proposed a spatio-temporal compression method to aggregate individual events into a few time steps of synaptic current to reduce the training and inference latency. To keep the accuracy of SNNs under high compression ratios, we also proposed a synaptic convolutional block to balance the dramatic changes between adjacent time steps. And multi-threshold Leaky Integrate-and-Fire (LIF) models with learnable membrane time constants are introduced to increase their information processing capability. We evaluate the proposed method for event stream classification tasks on neuromorphic N-MNIST, CIFAR10-DVS, and DVS128 gesture datasets. The experiment results show that our proposed method outperforms the state-of-the-art accuracy on nearly all datasets, using fewer time steps.

Development of Cement Stabilised Compressed Blocks Using Coal Bottom Ash: Influence of the Grain Size on Mechanical and Physical Properties of Blocks

This study investigated the physical and mechanical properties of cement-stabilized compressed blocks manufactured with coal bottom ash sourced from a power plant in Niger. Three different grain sizes were used for the production of compressed blocks with a hand-operated press. Thermal, hydric, mechanical, and fire resistance properties were assessed on the samples. It was found that the use of finer bottom ash resulted in lighter blocks with a density of about 1.02 mg/m3 and thermal conductivity in the range of 0.27 – 0.41 W/m·K. The size of the bottom ash used for the production of blocks did not significantly affect the value of mechanical strength. The exposure of blocks to temperatures of 200°C and 400°C did not reduce the strength of the samples. Neat bottom ash blocks can offer better thermal properties than typical building materials and provide acceptable mechanical strength.

Nonlinear compression block codes: Exact and random search strategy

The nonlinear block codes have significantly better testing abilities than the linear ones used in Built-In Self Test Equipment or test pattern decompressors. Unfortunately, no sufficiently effective deterministic code creation strategy has been discovered till now. This paper deals with a deterministic nonlinear code search strategy and a strategy based on extending linear compression codes by nonlinear check bits that improve the usability of decompressed patterns for testing circuits with more inputs. The earlier works used a purely or partially random search of the nonlinear check-bits truth tables to construct the first nonlinear structures. Here, we derive deterministic rules that can be used for the iterative process of code generation with no necessity of any random search for codes with the number of specified bits equal to three. We compare the code parameters of codes based on linear code extension, direct deterministic code search methods, and previously published results. The most extensive codes still need to be constructed with a partial random search strategy, but keeping the restrictions introduced in the paper causes the search to be substantially more efficient.

Kronecker CP Decomposition With Fast Multiplication for Compressing RNNs.

Recurrent neural networks (RNNs) are powerful in the tasks oriented to sequential data, such as natural language processing and video recognition. However, because the modern RNNs have complex topologies and expensive space/computation complexity, compressing them becomes a hot and promising topic in recent years. Among plenty of compression methods, tensor decomposition, e.g., tensor train (TT), block term (BT), tensor ring (TR), and hierarchical Tucker (HT), appears to be the most amazing approach because a very high compression ratio might be obtained. Nevertheless, none of these tensor decomposition formats can provide both space and computation efficiency. In this article, we consider to compress RNNs based on a novel Kronecker CANDECOMP/PARAFAC (KCP) decomposition, which is derived from Kronecker tensor (KT) decomposition, by proposing two fast algorithms of multiplication between the input and the tensor-decomposed weight. According to our experiments based on UCF11, Youtube Celebrities Face, UCF50, TIMIT, TED-LIUM, and Spiking Heidelberg digits datasets, it can be verified that the proposed KCP-RNNs have a comparable performance of accuracy with those in other tensor-decomposed formats, and even 278 219× compression ratio could be obtained by the low-rank KCP. More importantly, KCP-RNNs are efficient in both space and computation complexity compared with other tensor-decomposed ones. Besides, we find KCP has the best potential of parallel computing to accelerate the calculations in neural networks.

Erratum to: Bonding Strength between Cross Layers of CLT Evaluated by Torsion and Compression Block Shear Tests [Mokuzai Gakkaishi Vol.67 (2021) No.2 p.100-108

Erratum to: Bonding Strength between Cross Layers of CLT Evaluated by Torsion and Compression Block Shear Tests [Mokuzai Gakkaishi Vol.67 (2021) No.2 p.100-108

A review on addition of egg shell powder to CSEB blocks

Dry stack masonry units known as “Compressed Stabilized Earth Blocks” (CSEB) are constructed using dirt and often stabilised with cement. It is also known as a pressed earth block or a compressed soil block. The rise of a movement against environment today has awakened us to adapt to and incorporate practices that are pro-earth. Construction using CSEBs are one of the many ways we can reduce negative impact on earth. It is essential to use sustainable building materials in place of cement since cement-concrete buildings use roughly half of the energy produced and are responsible for significant CO2 emissions. Similarly addition of Egg Shell Powder (ESP) to CSEB blocks can provide binding properties to the block as it is rich in calcium (CaCo3), partial replacement of Egg Shell Powder (ESP) to cement can be effective in reducing environmental pollution.The experimental investigation are carried out to study the effect of Egg Shell Powder in Compressed Stabilized Earth Blocks by partial replacement with cement.It also provides structural strength and are biodegradable, sound proof and resistance to fire and pests.To introduce a concrete by adding egg shell powder with different levels of percentages. “To perform and compare the compression strength of CSEB blocks with and without replacement of cement. Egg shell powder possesses similar properties to that of cement”.

A reassessment of the CSA S304 chi-factor for design of reinforced concrete masonry members: large-scale experimental investigation

An experimental investigation was conducted that consisted of 12 pairs of reinforced masonry specimens in which the two specimens were constructed with their longitudinal axes oriented parallel and perpendicular to the laboratory floor, respectively. This work was conducted to reassess the need for the χ-factor accounting for the direction of compressive stresses in masonry members in relation to the direction used to establish their masonry assemblage strength included in CSA S304-14— Design of Masonry Structures. Parameters investigated included block size and reinforcement ratio. Test results suggest that the χ-factor can be eliminated from future editions of CSA S304 but should be considered with caution: a review of the literature and the current tests show that, in some cases, voids may develop between the exterior webs of adjacent blocks and so may be the source of the reduction in the magnitude of the equivalent rectangular compressive stress block.

Estimating the flexural strength of corroded reinforced concrete beams based on rectangular compressive stress block

Estimating the flexural strength of corroded reinforced concrete beams based on rectangular compressive stress block

Simplified shear equation of deep concrete beam considering orientation effect of opening and mechanical properties of fibre-concrete interface

In the present era of technology, the design of structural deep concrete beams is highly modernized through the use of computer-aided tools. However, theories for such design are the paramount aspects to be understood particularly when the beam has openings. To improve the mechanical properties of the deep beam with openings, recycled tyre steel fibres are required. To estimate the bond strength of the fibre-concrete interface, the concrete shrinkage strains for fibre lengths 30, 50 and 60 mm with a content of 0.5% were considered. To mitigate the conservativeness of some available shear models and improve the design of deep beams, the simplified shear equation model was developed. The model was established using a simplified compressive stress block, forces in steel reinforcement and shear stress at the fibre-concrete interface. The combined effect of opening height and length was also considered in the model. The results show that incorporating fibres in concrete increase the shear performance of deep concrete beams with openings due to high strains in the shear zone indicating high loads being transferred. For instance, beam BS2 had a strain of 0.0153 in the lower load path compared to 5.6 × 10−5 for beam BC2 There was a good correlation between measured and proposed shear capacities with t-test values of 0.46, 0.996 and 0.003 for beams without fibres, and with fibres and mesh respectively. Results also showed the model shear equation performed better compared to other equations with mean absolute error (MAE) and coefficient of variation (COV) of 9.3 and 18.9%, respectively for the control beams with openings. The model also showed a mean absolute error (MAE) and coefficient of variation (COV) of 11.6 and 7.4%, respectively for the beams with fibres. The COV and MAE for the proposed model were small than those in the database, therefore, the proposed model can provide the precise design of deep concrete beams with openings and fibres.

BTS-ST: Swin transformer network for segmentation and classification of multimodality breast cancer images

Breast cancer is considered the most commonly diagnosed cancer globally and falls second to lung cancer. For the early detection of breast tumors in women, breast cancer analysis using Ultrasound, Mammography, and MRI modalities as the initial screening process. Due to the random variation, irregular shapes, and blurred boundaries of tumor regions, the accurate segmentation of breast tumors is still a tricky task. The existing convolutional neural networks (CNNs) inherit their limitation by extracting global context information and, in most cases, proved less efficient in obtaining satisfactory results. As a solution, we proposed the BTS-ST network, a novel solution for breast tumor segmentation and classification that Swin-Transformer (ST) inspires. The BTS-ST network incorporates Swin-Transformer into traditional CNNs-based U-Net to improve global modeling capabilities. To improve the feature representation capability of irregularly shaped tumors, we first introduced a Spatial Interaction block (SIB), encoding spatial knowledge in the Swin Transformer block by developing pixel-level correlation. The segmentation accuracy of small-scale tumor regions is increased by building a Feature Compression block (FCB) to prevent information loss and compress smaller-scale features in patch token down sampling of Swin-Transformer. Finally, a Relationship Aggregation block (RAB) is developed as a bridge between dual encoders to combine global dependencies from Swin-Transformer into the features from CNN hierarchically. Extensive experiments are performed on breast tumor segmentation and classification tasks using multimodality Ultrasound, Mammogram, and MRI-based datasets. The results demonstrate that our proposed solution is comparatively better than other state-of-the-art methods.

A Study of the Behaviour of Concrete using Rice Husk Ash

Pollution waste is regularly increasing, and environmental preservation continues to grow a complex issue owing to the manufacture of burnt bricks & poor management of waste. This study uses an unfired compressed dirt block stabilised by solid waste as the test subject. Tests here on cubes was carried out to assess their compression characteristics & mechanical characteristics. 2 type of samples are prepared one with regulated Rice husk temperature burn (500-650 c) & another one is unregulated rice husk temperature burnt. Total number of blocks created are 36 blocks. The sample is prepared of different amounts of R-h replacement 10percent, 15 percent, 20 percent & 30 percent; R-H Ash with 20 percent replacement of regulated one have better mechanical properties overall. These investigation results revealed that blocks and the right proportion of waste products might be made. The mechanical properties of a block containing R-h ash are also increased.

An application of deep dual convolutional neural network for enhanced medical image denoising.

This work investigates the medical image denoising (MID) application of the dual denoising network (DudeNet) model for chest X-ray (CXR). The DudeNet model comprises four components: a feature extraction block with a sparse mechanism, an enhancement block, a compression block, and a reconstruction block. The developed model uses residual learning to boost denoising performance and batch normalization to accelerate the training process. The name proposed for this model is dual convolutional medical image-enhanced denoising network (DCMIEDNet). The peak signal-to-noise ratio (PSNR) and structure similarity index measurement (SSIM) are used to assess the MID performance for five different additive white Gaussian noise (AWGN) levels of σ = 15, 25, 40, 50, and 60 in CXR images. Presented investigations revealed that the PSNR and SSIM offered by DCMIEDNet are better than several popular state-of-the-art models such as block matching and 3D filtering, denoising convolutional neural network, and feature-guided denoising convolutional neural network. In addition, it is also superior to the recently reported MID models like deep convolutional neural network with residual learning, real-valued medical image denoising network, and complex-valued medical image denoising network. Therefore, based on the presented experiments, it is concluded that applying the DudeNet methodology for DCMIEDNet promises to be quite helpful for physicians.

Experimental Study on the Production and Mechanical Behavior of Compressed Lime-Cement-Stabilized Interlock Soil Blocks

Compressed stabilized soil block is a sustainable building material primarily made up of stabilized damp soil compressed under pressure. Soil properties and the type of the stabilizer used in producing compressed soil blocks have a significant impact on the quality and behavior of the soil blocks. This study presents the physical and mechanical behavior of lime-cement-stabilized compressed interlock soil blocks produced from two types of natural soil. The two types of soil have different index properties and mineral oxide compositions. Lime-cement combination and cement standalone was used as a binder in the production of test sample blocks depending on the index properties of the soil. 2%lime + 6%cement, 3%lime + 8%cement, and 4%lime + 10%cement were used for the soil block produced from silty clay soil of medium plasticity index. On the other hand, 6%, 8%, and 10% cement by dry mass of soil were used to stabilize silty sand soil. The behaviors of the blocks, such as dry density, the initial rate of water absorption, saturated absorption of water, compressive strength, and stress-strain relation, were examined. The result shows that the compressed soil blocks produced from lime-cement-stabilized silty clay soil has a low rate of initial water absorption and a low dry unit weight when compared to cement-stabilized sandy soil blocks. Soil blocks produced from cement-stabilized silty sand soil attain greater compressive strength by more than 50% of the compressive strength of silty clay soil blocks stabilized by a combination of lime and cement at 60 days after production. The initial tangent modulus of the soil blocks produced using a manual compressing machine from a clay soil stabilized by the lime-cement proportions of 2%L + 6%C, 3%L + 8%C, and 4%L + 10%C is about 1,700 MPa–2,300 MPa with a dry density greater than 1,660 kg/m3.

Tailoring Dense, Orientation-Tunable, and Interleavedly Structured Carbon-Based Heat Dissipation Plates.

The controllability of the microstructure of a compressed hierarchical building block is essential for optimizing a variety of performance parameters, such as thermal management. However, owing to the strong orientation effect during compression molding, optimizing the alignment of materials perpendicular to the direction of pressure is challenging. Herein, to illustrate the effect of the ordered microstructure on heat dissipation, thermally conductive carbon-based materials are fabricated by tailoring dense, orientation-tunable, and interleaved structures. Vertically aligned carbon nanotube arrays (VACNTs) interconnected with graphene films (GF) are prepared as a 3D core-ordered material to fabricate compressed building blocks of O-VA-GF and S-VA-GF. Leveraging the densified interleaved structure offered by VACNTs, the hierarchical O-VA-GF achieves excellent through-plane (41.7W m-1 K-1 ) and in-plane (397.9W m-1 K-1 ) thermal conductivities, outperforming similar composites of S-VA-GF (through-plane: 10.3W m-1 K-1 and in-plane: 240.9W m-1 K-1 ) with horizontally collapsed carbon nanotubes. As heat dissipation plates, these orderly assembled composites yield a 144% and 44% enhancement in the cooling coefficient compared with conventional Si3 N4 for cooling high-power light-emitting diode chips.