Multi-scale Structural Features Research Articles

Improving the mechanical and thermal performance of bio-based concrete through multi-objective optimization

Bio-based concrete, which utilizes recycled agricultural waste as aggregates, has emerged as a sustainable alternative material in the construction industry. Optimizing its combined mechanical and thermal performance can promote its environmental and energy benefits. To achieve this, this paper presents a multi-objective optimization approach to improve the overall performance of bio-based concrete based on multiscale structural features, considering the porosity, shape, orientation, and volume fraction. Optimization strategies are provided for three main industrial applications. Furthermore, this study addresses multiple design requirements, including high-volume bio-aggregates, more demanding performance requirements, low densities, and humid environments. The optimization results demonstrate that the optimal solution varies with industrial applications and requirements. Consequently, we quantitatively provide specific optimal solutions for each case, and offer designers three preferences for mechanical properties, insulation, or balanced performance. Overall, this study optimizes the comprehensive performance of bio-based concrete in terms of multiscale structures, thereby contributing to a more sustainable construction industry.

Structure-Aware Cross-Modal Transformer for Depth Completion.

In this paper, we present a Structure-aware Cross-Modal Transformer (SCMT) to fully capture the 3D structures hidden in sparse depths for depth completion. Most existing methods learn to predict dense depths by taking depths as an additional channel of RGB images or learning 2D affinities to perform depth propagation. However, they fail to exploit 3D structures implied in the depth channel, thereby losing the informative 3D knowledge that provides important priors to distinguish the foreground and background features. Moreover, since these methods rely on the color textures of 2D images, it is challenging for them to handle poor-texture regions without the guidance of explicit 3D cues. To address this, we disentangle the hierarchical 3D scene-level structure from the RGB-D input and construct a pathway to make sharp depth boundaries and object shape outlines accessible to 2D features. Specifically, we extract 2D and 3D features from depth inputs and the back-projected point clouds respectively by building a two-stream network. To leverage 3D structures, we construct several cross-modal transformers to adaptively propagate multi-scale 3D structural features to the 2D stream, energizing 2D features with priors of object shapes and local geometries. Experimental results show that our SCMT achieves state-of-the-art performance on three popular outdoor (KITTI) and indoor (VOID and NYU) benchmarks.

Deep Learning Model for Quality Assessment of Urinary Bladder Ultrasound Images using Multi-scale and Higher-order Processing.

Autonomous Ultrasound Image Quality Assessment (US-IQA) is a promising tool to aid the interpretation by practicing sonographers and to enable the future robotization of ultrasound procedures. However, autonomous US-IQA has several challenges. Ultrasound images contain many spurious artifacts, such as noise due to handheld probe positioning, errors in the selection of probe parameters and patient respiration during the procedure. Further, these images are highly variable in appearance with respect to the individual patient's physiology. We propose to use a deep Convolutional Neural Network (CNN), USQNet, which utilizes a Multi-scale and Local-to-Global Second-order Pooling (MS-L2GSoP) classifier to conduct the sonographer-like assessment of image quality. This classifier first extracts features at multiple scales to encode the inter-patient anatomical variations, similar to a sonographer's understanding of anatomy. Then, it uses second-order pooling in the intermediate layers (local) and at the end of the network (global) to exploit the second-order statistical dependency of multi-scale structural and multi-region textural features. The L2GSoP will capture the higher-order relationships between different spatial locations and provide the seed for correlating local patches, much like a sonographer prioritizes regions across the image. We experimentally validated the USQNet for a new dataset of the human urinary bladder ultrasound images. The validation involved first with the subjective assessment by experienced radiologists' annotation, and then with state-of-the-art CNN networks for US-IQA and its ablated counterparts. The results demonstrate that USQNet achieves a remarkable accuracy of 92.4% and outperforms the SOTA models by 3 - 14% while requiring comparable computation time.

Polymer Bead Foams: A Review on Foam Preparation, Molding, and Interbead Bonding Mechanism

The diverse physical appearances and wide density range of polymer bead foams offer immense potential in various applications and future advancements. The multiscale and multilevel structural features of bead foams involve many fundamental scientific topics. This review presents a comprehensive overview of recent progress in the preparation and molding techniques of bead foams. Firstly, it gives a comparative analysis on the bead foam characteristics of distinct polymers. Then, a summary and comparison of molding techniques employed for fabricating bead foam parts are provided. Beyond traditional methods like steam-chest molding (SCM) and adhesive-assisted molding (AAM), emerging techniques like in-mold foaming and molding (IMFM) and microwave selective sintering (MSS) are highlighted. Lastly, the bonding mechanisms behind these diverse molding methods are discussed.

MsCNN-LSTM perimeter intrusion vibration signal identification method based on ultra-weak FBG arrays

To improve the recognition rate of ultra-weak fiber Bragg grating (UWFBG) arrays in perimeter security monitoring, this paper proposes a msCNN-LSTM combinatorial model recognition method, which combines an improved multi-scale convolutional neural network (msCNN) and a long short-term memory neural network (LSTM) to identify intrusion behaviors more accurately by synchronously extracting the multi-scale structural features and time-dependent relationships of vibration signals. Using msCNN to cross-learn multi-layer features, the hidden information between features at different scales can be obtained. The attention mechanism was applied to recalibrate the combined features extracted from the shallower layers, which enhanced the expression ability of features. Finally, the time dependence is analyzed by LSTM. Experimental results show that the scheme can effectively distinguish five typical intrusion behaviors, and the average recognition rate can reach 97.84%.

Unleashing multi-scale mechanical enhancement in NiTi shape memory alloy via annular intra-laser deposition with homogenized Ti2Ni nanoprecipitates

Additive manufacturing (AM) of composition-sensitive NiTi shape memory alloys (SMAs) is hindered by uncontrollable non-equilibrium microstructures, leading to functional performance degradation and limitations in high-throughput net-shaping applications. To address these challenges, we propose a novel approach combining annular intra-laser directed energy deposition (AIL-DED) with a tailored gradient post-treatment process. This enables the fabrication of nanocomposite microstructures in NiTi SMAs with precise control over the homogenization of Ti2Ni nanoprecipitates. In-situ monitoring and characterization of thermal history, microstrain, and constitutional phases, were utilized to uncover the relationship between multi-scale structural features and mechanical responses. The interplay between laser reflectivity, energy input, and lattice thermal vibrations affects forming quality in AIL-DED. Dynamic temperature variations were studied before reaching thermal equilibrium, revealing insights into thermal and mechanical interactions during the process. These findings contribute to understanding challenges in fabricating high-quality NiTi SMAs. Examining phase transformation behavior and its correlation with processing parameters deepens understanding of microstructural evolution and provides insights for tailoring material properties. The stress relaxation and dislocation evacuation at elevated temperatures caused the expansion of lattice and interplanar spacing in the age-treated SMA's B2 matrix phase, exhibiting a stable two-step phase transformation behavior (B19′↔R↔B2). AIL-DED combined with two-stage heat treatment strategy demonstrated enhanced homogeneity of non-equilibrium microstructures, resulting in the formation of high-density, well-dispersed Ti2Ni nanoprecipitates, with a semi-coherent interface to B2 matrix. The combination of load-bearing strengthening through well-dispersed Ti2Ni nanoprecipitates and Orowan strengthening at the semi-coherent interface yielded simultaneous enhancements in nanohardness, ultimate compressive strength, compressive fracture strain, and deformation recovery, following the AIL-DED combined with two-stage heat treatment process. The fracture mechanism transitioned from a mixture of brittle-ductile behavior to predominantly ductile characteristics. This study provides valuable insights that advance additive manufacturing and enhance the functional performance of SMAs in critical applications.

Micromechanical homogenization of a hydrogel-filled electrospun scaffold for tissue-engineered epicardial patching of the infarcted heart: a feasibility study

AbstractFor tissue engineering applications, accurate prediction of the effective mechanical properties of tissue scaffolds is critical. Open and closed cell modelling, mean-field homogenization theory, and finite element (FE) methods are theories and techniques currently used in conventional homogenization methods to estimate the equivalent mechanical properties of tissue-engineering scaffolds. This study aimed at developing a formulation to link the microscopic structure and macroscopic mechanics of a fibrous electrospun scaffold filled with a hydrogel for use as an epicardial patch for local support of the infarcted heart. The macroscopic elastic modulus of the scaffold was predicted to be 0.287 MPa with the FE method and 0.290 MPa with the closed-cell model for the realistic fibre structure of the scaffold, and 0.108 MPa and 0.540 MPa with mean-field homogenization for randomly oriented and completely aligned fibres. The homogenized constitutive description of the scaffold was implemented for an epicardial patch in a FE model of a human cardiac left ventricle to assess the effects of patching on myocardial mechanics and ventricular function in the presence of an infarct. Epicardial patching was predicted to reduce maximum myocardial stress in the infarcted LV from 19 kPa (no patch) to 9.5 kPa (patch) and to marginally improve the ventricular ejection fraction from 40% (no patch) to 43% (patch). This study demonstrates the feasibility of homogenization techniques to represent complex multiscale structural features in a simplified but meaningful and effective manner.

Recent Experimental Advances in Characterizing the Self-Assembly and Phase Behavior of Polypeptoids.

Polypeptoids are a family of synthetic peptidomimetic polymers featuring N-substituted polyglycine backbones with large chemical and structural diversity. Their synthetic accessibility, tunable property/functionality, and biological relevance make polypeptoids a promising platform for molecular biomimicry and various biotechnological applications. To gain insight into the relationship between the chemical structure, self-assembly behavior, and physicochemical properties of polypeptoids, many efforts have been made using thermal analysis, microscopy, scattering, and spectroscopic techniques. In this review, we summarize recent experimental investigations that have focused on the hierarchical self-assembly and phase behavior of polypeptoids in bulk, thin film, and solution states, highlighting the use of advanced characterization tools such as in situ microscopy and scattering techniques. These methods enable researchers to unravel multiscale structural features and assembly processes of polypeptoids over a wide range of length and time scales, thereby providing new insights into the structure-property relationship of these protein-mimetic materials.

Different multi-scale structural features of oat resistant starch prepared by ultrasound combined enzymatic hydrolysis affect its digestive properties

In this research, oat resistant starch (ORS) was prepared by autoclaving-retrogradation cycle (ORS-A), enzymatic hydrolysis (ORS-B), and ultrasound combined enzymatic hydrolysis (ORS-C). Differences in their structural features, physicochemical properties and digestive properties were studied. Results of particle size distribution, XRD, DSC, FTIR, SEM and in vitro digestion showed that ORS-C was a B + C-crystal, and ORS-C had a larger particle size, the smallest span value, the highest relative crystallinity, the most ordered and stable double helix structure, the roughest surface shape and strongest digestion resistance compared to ORS-A and ORS-B. Correlation analysis revealed that the digestion resistance of ORS-C was strongly positively correlated with RS content, amylose content, relative crystallinity and absorption peak intensity ratio of 1047/1022 cm−1 (R1047/1022), and weakly positively correlated with average particle size. These results provided theoretical support for the application of ORS-C with strong digestion resistance prepared by ultrasound combined enzymatic hydrolysis in the low GI food application.

Stacked graph bone region U-net with bone representation for hand pose estimation and semi-supervised training

3D hand estimation from 2D joint information is an essential task in human-machine interaction, which has achieved great progress as an application of deep learning. However, regression-based methods do not perform well because the structural information is not effectively exploited, and the joint coordinates are variable. To address these issues, the hand pose is represented with bone vectors instead of joint coordinates in this study, which are stabler to learn and allow for easier encoding of the hand geometric structure and joint dependency. A novel graph bone region U-Net is specifically designed for bone representation to learn multiscale structural features, where the proposed novel elements (graph convolution, pooling and unpooling) incorporate hand structural knowledge. Under the introduced “finger-to-hand” framework, the network gradually decreases the scale from bone to finger to hand for learning more meaningful multiscale features. Moreover, the unit network is stacked repeatedly to extract multilevel features. Based on the above network, a simple but effective semi-supervised approach is introduced to address the lack of 3D hand pose labels. Many experiments are conducted to evaluate the proposed approach on two challenging datasets. The experimental results show that the proposed supervised approach outperforms the state-of-the-art methods, and the proposed semi-supervised approach can still achieve favorable performance when the labeled data are scarce.

GRPAFusion: A Gradient Residual and Pyramid Attention-Based Multiscale Network for Multimodal Image Fusion

Multimodal image fusion aims to retain valid information from different modalities, remove redundant information to highlight critical targets, and maintain rich texture details in the fused image. However, current image fusion networks only use simple convolutional layers to extract features, ignoring global dependencies and channel contexts. This paper proposes GRPAFusion, a multimodal image fusion framework based on gradient residual and pyramid attention. The framework uses multiscale gradient residual blocks to extract multiscale structural features and multigranularity detail features from the source image. The depth features from different modalities were adaptively corrected for inter-channel responses using a pyramid split attention module to generate high-quality fused images. Experimental results on public datasets indicated that GRPAFusion outperforms the current fusion methods in subjective and objective evaluations.

Remote Sensing Scene Image Classification Based on mmsCNN–HMM with Stacking Ensemble Model

The development of convolution neural networks (CNNs) has become a significant means to solve the problem of remote sensing scene image classification. However, well-performing CNNs generally have high complexity and are prone to overfitting. To handle the above problem, we present a new classification approach using an mmsCNN–HMM combined model with stacking ensemble mechanism in this paper. First of all, a modified multi-scale convolution neural network (mmsCNN) is proposed to extract multi-scale structural features, which has a lightweight structure and can avoid high computational complexity. Then, we utilize a hidden Markov model (HMM) to mine the context information of the extracted features of the whole sample image. For different categories of scene images, the corresponding HMM is trained and all the trained HMMs form an HMM group. In addition, our approach is based on a stacking ensemble learning scheme, in which the preliminary predicted values generated by the HMM group are used in an extreme gradient boosting (XGBoost) model to generate the final prediction. This stacking ensemble learning mechanism integrates multiple models to make decisions together, which can effectively prevent overfitting while ensuring accuracy. Finally, the trained XGBoost model conducts the scene category prediction. In this paper, the six most widely used remote sensing scene datasets, UCM, RSSCN, SIRI-WHU, WHU-RS, AID, and NWPU, are selected to carry out all kinds of experiments. The numerical experiments verify that the proposed approach shows more important advantages than the advanced approaches.

Pea-Resistant Starch with Different Multi-scale Structural Features Attenuates the Obesity-Related Physiological Changes in High-Fat Diet Mice.

The present study compared the modulatory effects of different resistant starches (RSs) isolated from native (NP-RS), acid-hydrolyzed (AHP-RS), and pullulanase debranched (PDP-RS) pea starches on the corresponding in vivo metabolic responses in high fat (HF)-diet-induced obese mice. The biochemical studies on serum lipid profile and antioxidant enzyme activities were supported by histological and gene expression analyses, which suggested a potential therapeutic role for RS in regulating obesity, possibly through the production of short-chain fatty acids and the proliferation of some beneficial colonic bacteria, including Allobaculum, Bifidobacterium, Odoribacter, Clostridium, and Prevotella. Particularly, a more pronounced effect of AHP-RS with a higher proportion of the crystalline region and a more ordered double-helical alignment on improving the hyperlipidemic symptoms in obese mice induced by a HF diet was observed. Our analysis revealed that the RS3 samples seemed to be more effective than RS2 in terms of attenuating obesity in mice that were fed a HF diet.

Freeze-derived heterogeneous structural color films

Structural colors have a demonstrated value in constructing various functional materials. Efforts in this area are devoted to developing stratagem for generating heterogeneous structurally colored materials with new architectures and functions. Here, inspired by icing process in nature and ice-templating technologies, we present freeze-derived heterogeneous structural color hydrogels with multiscale structural and functional features. We find that the space-occupying effect of ice crystals is helpful for tuning the distance of non-close-packed colloidal crystal nanoparticles, resulting in corresponding reflection wavelength shifts in the icing area. Thus, by effectively controlling the growth of ice crystals and photo-polymerizing them, structural color hydrogels with the desired structures and morphologies can be customized. Other than traditional monochromatic structure color hydrogels, the resultant hydrogels can be imparted with heterogeneous structured multi-compartment body and multi-color with designed patterns through varying the freezing area design. Based on these features, we have also explored the potential value of these heterotypic structural color hydrogels for information encryptions and decryptions by creating spatiotemporally controlled icing areas. We believe that these inverse ice-template structural color hydrogels will offer new routes for the construction and modulation of next generation smart materials with desired complex architectures.

Single Image Dehazing Based on Two-Stream Convolutional Neural Network

Objective The haze weather environment leads to the deterioration of the visual effect of the image, and it is difficult to carry out the work of the advanced vision task. Therefore, dehazing the haze image is an important step before the execution of the advanced vision task. Traditional dehazing algorithms achieve image dehazing by improving image brightness and contrast or constructing artificial priors such as color attenuation priors and dark channel priors, but the effect is unstable when dealing with complex scenes. In the method based on convolutional neural network, the image dehazing network of the encoding and decoding structure does not consider the difference before and after the dehazing image, and the image spatial information is lost in the encoding stage. In order to overcome these problems, this paper proposes a novel end-to-end two-stream convolutional neural network for single image dehazing. Method The network model is composed of a spatial information feature stream and a high-level semantic feature stream. The spatial information feature stream retains the detailed information of the dehazing image, and the high-level semantic feature stream extracts the multi-scale structural features of the dehazing image. A spatial information auxiliary module is designed between the feature streams. This module uses the attention mechanism to construct a unified expression of different types of information, and realizes the gradual restoration of the clear image with the semantic information auxiliary spatial information in the dehazing network. A parallel residual twicing module is proposed, which performs dehazing on the difference information of features at different stages to improve the model’s ability to discriminate haze images. Result The peak signal-to-noise ratio and structural similarity are used to quantitatively evaluate the similarity between the dehazing results of each algorithm and the original image. The structure similarity and peak signal-to-noise ratio of the method in this paper reached 0.852 and 17.557dB on the Hazerd dataset, which were higher than all comparison algorithms. On the SOTS dataset, the indicators are 0.955 and 27.348dB, which are sub-optimal results. In experiments with real haze images, this method can also achieve excellent visual restoration effects.

SFOC: A NOVEL MULTI-DIRECTIONAL AND MULTI-SCALE STRUCTURAL DESCRIPTOR FOR MULTIMODAL REMOTE SENSING IMAGE MATCHING

Abstract. Accurate matching of multimodal remote sensing (RS) images (e.g., optical, infrared, LiDAR, SAR, and rasterized maps) is still an ongoing challenge because of nonlinear radiometric differences (NRD) between these images. Considering that structural properties are preserved between multimodal images, this paper proposes a robust matching method based on multi-directional and multi-scale structural features, which consist of two critical steps. Firstly, a novel structural descriptor named the Steerable Filters of first- and second-Order Channels (SFOC) is constructed to address severe NRD, which combines the first- and second-order gradient information by using the steerable filters to depict multi-directional and multi-scale structural features of images. Meanwhile, SFOC is further enhanced by performing the dilated Gaussian convolutions with different dilated rates on it, which can capture multi-level context structural features and improve the ability to resist noise. Then, a fast similarity measure, called Fast Normalized Cross-Correlation (Fast-NCCSFOC), is established to detect correspondences by a template matching scheme, which employs the Fast Fourier Transform (FFT) technique and the integral image to improve the matching efficiency. The performance of the proposed SFOC has been evaluated with many different kinds of multimodal RS images, and experimental results show its superior matching performance compared with the state-of-the-art methods.

Multiscale DenseNet Meets With Bi-RNN for Hyperspectral Image Classification

Convolutional neural network (CNN) has been successfully introduced to hyperspectral image (HSI) classification and achieved effective performance. With the depth of the CNN increases, it may cause the gradient to become zero, and the structure lacks the utilization of the correlated spatial feature information between different convolutional layers. At the same time, this single-scale convolution kernel is insufficient in expressing the complex spatial structure information of HSI. Additionally, the CNN-based methods treat the HSIs spectral band data as a disordered vector in the process of feature extraction, which abandons the exploitation of its internal spectral correlations. To address these issues, we propose a novel spectral-spatial network classification framework based on multi-scale dense connected convolutional network (DenseNet) and bi-direction recurrent neural network (Bi-RNN) with attention mechanism network (MDRN). For the proposed MDRN, in terms of spatial feature extraction, a multi-scale DenseNet is exploited to combine shallow and deep convolution features to extract the multi-scale and complex spatial structure features at each layer. In the aspects of spectral feature extraction, Bi-RNN with attention mechanism is used to capture the inner spectral correlations within a continuous spectrum. Three standard real hyperspectral datasets were used to verify the effectiveness of the proposed MDRN approach. Experimental results indicate that the proposed MDRN method can make full use of the spectral and spatial information of the image, and it has better performance than some advanced algorithms in HSI classification. Finally, in the application of hyperspectral data captured by Gaofen-5 (GF-5) satellite, the practicability of the proposed MDRN method is also superior to other methods.

Multiscale structural mapping of Alzheimer’s disease neurodegeneration

The recently described biological framework of Alzheimer’s disease (AD) emphasizes three types of pathology to characterize this disorder, referred to as the ‘amyloid/tau/neurodegeneration’ (A-T-N) status. The ‘neurodegenerative’ component is typically defined by atrophy measures derived from structural magnetic resonance imaging (MRI) such as hippocampal volume. Neurodegeneration measures from imaging are associated with disease symptoms and prognosis. Thus, sensitive image-based quantification of neurodegeneration in AD has an important role in a range of clinical and research operations. Although hippocampal volume is a sensitive metric of neurodegeneration, this measure is impacted by several clinical conditions other than AD and therefore lacks specificity. In contrast, selective regional cortical atrophy, known as the ‘cortical signature of AD’ provides greater specificity to AD pathology. Although atrophy is apparent even in the preclinical stages of the disease, it is possible that increased sensitivity to degeneration could be achieved by including tissue microstructural properties in the neurodegeneration measure. However, to facilitate clinical feasibility, such information should be obtainable from a single, short, noninvasive imaging protocol. We propose a multiscale MRI procedure that advances prior work through the quantification of features at both macrostructural (morphometry) and microstructural (tissue properties obtained from multiple layers of cortex and subcortical white matter) scales from a single structural brain image (referred to as ‘multi-scale structural mapping’; MSSM). Vertex-wise partial least squares (PLS) regression was used to compress these multi-scale structural features. When contrasting patients with AD to cognitively intact matched older adults, the MSSM procedure showed substantially broader regional group differences including areas that were not statistically significant when using cortical thickness alone. Further, with multiple machine learning algorithms and ensemble procedures, we found that MSSM provides accurate detection of individuals with AD dementia (AUROC = 0.962, AUPRC = 0.976) and individuals with mild cognitive impairment (MCI) that subsequently progressed to AD dementia (AUROC = 0.908, AUPRC = 0.910). The findings demonstrate the critical advancement of neurodegeneration quantification provided through multiscale mapping. Future work will determine the sensitivity of this technique for accurately detecting individuals with earlier impairment and biomarker positivity in the absence of impairment.

Detection of excavated areas in high-resolution remote sensing imagery using combined hierarchical spatial pyramid pooling and VGGNet

ABSTRACT The excavated areas formed by human activities often indicate the destruction of the ecological environment and topography, so it is of great significance to carry out dynamic monitoring. For the characteristics of the complex structure, different spatial scales, and less vegetation coverage in the excavated area, this study proposes a hierarchical spatial pyramid pooling (H-SPP) structure to effectively obtain the multiscale complex structural features from remote sensing images. Then, we add the H-SPP structure to the VGG-16 network to form a multiscale visual geometry group (M-VGG) suitable for excavated area detection and optimize the network by using the average pooling method to obtain more global information. The vegetation index is used to mask the background information. The experimental results show that the proposed method significantly improves the object detection performance for the excavated area compared with other methods. The detection time is reduced by almost 90%, and the accuracy is increased by more than 4%.

Simultaneous Extraction of Multi-Scale Structural Features and the Sequential Information With an End-To-End mCNN-HMM Combined Model for Fiber Distributed Acoustic Sensor

Presently, it is still challenging to obtain satisfied identification results for long-distance safety monitoring with fiber distributed acoustic sensor (DAS) in practical complicated burying environments. Thus, extracting increasingly abundant features has always been the direction of DAS signal recognition. This paper proposes a new recognition method using an end-to-end mCNN-HMM combined model, which can identify the vibration sources more correctly by simultaneously extracting multi-scale structural features and the sequential information of the DAS signals. A modified multi-scale convolution neural network (mCNN) is designed to automatically extract the DAS signals' local structural features from a multilevel perspective and their relationship in the proposed model. A hidden Markov model (HMM) is then used to mine the sequential information of the whole sample's previously extracted features. The test results based on real field data show that it outperforms the HMM model with the all-around hand-crafted features, the CNN-HMM model, and the MS-CNN-HMM model in both the feature extraction ability and the recognition accuracy in the case of little increase in time consumption. Moreover, the Euclidean distance between the posterior probabilities classified correctly and incorrectly is proposed to evaluate the test samples' feature distinguishability for different recognition models. Then the feature extraction capabilities of the models can be measured in an objective parameter.