Search

Abstract Purpose The new grading system for invasive nonmucinous lung adenocarcinoma (LUAD) in the 2021 World Health Organization Classification of Thoracic Tumors was based on a combination of histologically predominant subtypes and high-grade components. In this study, a model for the pretreatment prediction of grade 3 tumors was established according to new grading standards. Methods We retrospectively collected 399 cases of clinical stage I (cStage-I) LUAD surgically treated in Tianjin Chest Hospital from 2015 to 2018 as the training cohort. Besides, the validation cohort consists of 216 patients who were collected from 2019 to 2020. These patients were also diagnosed with clinical cStage-I LUAD and underwent surgical treatment at Tianjin Chest Hospital. Univariable and multivariable logistic regression analyses were used to select independent risk factors for grade 3 adenocarcinomas in the training cohort. The nomogram prediction model of grade 3 tumors was established by R software. Results In the training cohort, there were 155 grade 3 tumors (38.85%), the recurrence-free survival of which in the lobectomy subgroup was better than that in the sublobectomy subgroup (P = 0.034). After univariable and multivariable analysis, four predictors including consolidation-to-tumor ratio, CEA level, lobulation, and smoking history were incorporated into the model. A nomogram was established and internally validated by bootstrapping. The Hosmer–Lemeshow test result was χ2 = 7.052 (P = 0.531). The C-index and area under the receiver operating characteristic curve were 0.708 (95% CI: 0.6563–0.7586) for the training cohort and 0.713 (95% CI: 0.6426–0.7839) for the external validation cohort. Conclusions The nomogram prediction model of grade 3 LUAD was well fitted and can be used to assist in surgical or adjuvant treatment decision-making.

Abstract Objective To develop and validate a prediction model for early recurrence of stage I lung adenocarcinoma (LUAD) that combines radiomics features based on preoperative CT with tumour spread through air spaces (STAS). Materials and methods The most recent preoperative thin-section chest CT scans and postoperative pathological haematoxylin and eosin-stained sections were retrospectively collected from patients with a postoperative pathological diagnosis of stage I LUAD. Regions of interest were manually segmented, and radiomics features were extracted from the tumour and peritumoral regions extended by 3 voxel units, 6 voxel units, and 12 voxel units, and 2D and 3D deep learning image features were extracted by convolutional neural networks. Then, the RAdiomics Integrated with STAS model (RAISm) was constructed. The performance of RAISm was then evaluated in a development cohort and validation cohort. Results A total of 226 patients from two medical centres from January 2015 to December 2018 were retrospectively included as the development cohort for the model and were randomly split into a training set (72.6%, n = 164) and a test set (27.4%, n = 62). From June 2019 to December 2019, 51 patients were included in the validation cohort. RAISm had excellent discrimination in predicting the early recurrence of stage I LUAD in the training cohort (AUC = 0.847, 95% CI 0.762–0.932) and validation cohort (AUC = 0.817, 95% CI 0.625–1.000). RAISm outperformed single modality signatures and other combinations of signatures in terms of discrimination and clinical net benefits. Conclusion We pioneered combining preoperative CT-based radiomics with STAS to predict stage I LUAD recurrence postoperatively and confirmed the superior effect of the model in validation cohorts, showing its potential to assist in postoperative treatment strategies.

Abstract Background The current accuracy of frozen section diagnosis of tumor spread through air spaces (STAS) in non-small cell lung cancer (NSCLC) is poor. However, the accuracy and prognostic value of STAS assessment on frozen sections in small-sized NSCLC (diameter ≤ 2 cm) is unknown. Methods Three hundred fifty-two patients with clinical stage I NSCLC (≤ 2 cm) were included, of which the paraffin sections and frozen sections were reviewed. The accuracy of STAS diagnosis in frozen sections was assessed using paraffin sections as the gold standard. The relationship between STAS on frozen sections and prognosis was assessed by the Kaplan–Meier method and log-rank tests. Results STAS on frozen sections in 58 of 352 patients could not be evaluated. In the other 294 patients, 36.39% (107/294) was STAS-positive on paraffin sections and 29.59% (87/294) on frozen sections. The accuracy of frozen section diagnosis of STAS was 74.14% (218/294), sensitivity was 55.14% (59/107), specificity was 85.02% (159/187) and agreement was moderate (K = 0.418). In subgroup analysis, the Kappa values for frozen section diagnosis of STAS in the consolidation-to-tumor ratio (CTR) ≤ 0.5 group and CTR > 0.5 group were 0.368, 0.415, respectively. In survival analysis, STAS-positive frozen sections were associated with worse recurrence-free survival in the CTR > 0.5 group (P 0.5) suggests that frozen section assessment of STAS can be applied to the treatment strategy of small-sized NSCLC with CTR > 0.5.

Abstract Background The transmembrane receptor Kremen2 has been reported to participate in the tumorigenesis and metastasis of gastric cancer. However, the role of Kremen2 in non-small cell lung cancer (NSCLC) and the underlying mechanism remain unclear. This study aimed to explore the biological function and regulatory mechanism of Kremen2 in NSCLC. Methods The correlation between Kremen2 expression and NSCLC was assessed by analyzing the public database and clinical tissue samples. Colony formation and EdU assays were performed to examine cell proliferation. Transwell and wound healing assays were used to observe cell migration ability. Tumor-bearing nude mice and metastatic tumor models were used to detect the in vivo tumorigenic and metastatic abilities of the NSCLC cells. An immunohistochemical assay was used to detect the expression of proliferation-related proteins in tissues. Western blot, immunoprecipitation and immunofluorescence were conducted to elucidate the Kremen2 regulatory mechanisms in NSCLC. Results Kremen2 was highly expressed in tumor tissues from NSCLC patients and was positively correlated with a poor patient prognosis. Knockout or knockdown of Kremen2 inhibited cell proliferation and migration ability of NSCLC cells. In vivo knockdown of Kremen2 inhibited the tumorigenicity and number of metastatic nodules of NSCLC cells in nude mice. Mechanistically, Kremen2 interacted with suppressor of cytokine signaling 3 (SOCS3) to maintain the epidermal growth factor receptor (EGFR) protein levels by preventing SOCS3-mediated ubiquitination and degradation of EGFR, which, in turn, promoted activation of the PI3K-AKT and JAK2-STAT3 signaling pathways. Conclusions Our study identified Kremen2 as a candidate oncogene in NSCLC and may provide a potential target for NSCLC treatment.

Abstract Background Percutaneous needle biopsy (PNB) and bronchoscopic biopsy (BB) are widely used in the preoperative diagnosis of pulmonary nodules, but whether PNB or BB may cause tumor spread through air spaces (STAS) has not been reported. Methods 433 postoperative patients with pathological stage I non-small cell lung cancer (NSCLC) from January 2015 to December 2018 at our hospital were enrolled and divided into PNB group (n = 40), BB group (n = 48) and non-biopsy group (n = 345). The PNB and BB groups were matched using propensity score matched (PSM) separately from the non-biopsy group, after which the effects of PNB and BB on STAS, recurrence-free survival (RFS) and overall survival (OS) were assessed. Results After PSM for 9 confounding factors (gender, age, smoking history, tumor site, scope of surgery, pathology type, stage, maximum tumor diameter and postoperative treatment), 38 cases in the PNB group were successfully matched with 38 cases in the non-biopsy group and 28 cases in the BB group were successfully matched with 28 cases in the non-biopsy group. After PSM, there was no significant difference in the incidence of STAS between the PNB and non-biopsy groups (42.1% vs. 34.2%, P > 0.05) and between the BB and non-biopsy groups (42.9% vs. 46.4%, P > 0.05). The results after PSM showed no significant effect of both PNB and BB on RFS and OS after radical surgery (P > 0.05). Conclusion Preoperative biopsy in patients with stage I NSCLC has not been shown to increase the occurrence of STAS, nor postoperative recurrence and death.

BackgroundSmall-cell lung cancer (SCLC) is a recalcitrant tumor with a poor prognosis. With the rise of SCLC research in the past decade, this study aims to analyze the foundation and frontiers of SCLC research through bibliometric analysis.MethodsRelevant publications from the Web of Science Core Collection were retrieved on January 3, 2022. R package bibliometrix and EXCEL2019 were used to analyze quantitative variables. Bibliometric mapping was constructed by VOS viewer and CiteSpace software to visualize citation, co-authorship, co-occurrence, and co-citation analysis of countries/regions, organizations, authors, references, and keywords.ResultsA total of 2,361 publications related to SCLC were identified with the total amount of articles steadily increasing, where China is the most productive country with 859 papers. Scholars and organizations from the United States, China, and Europe are primary sources of this research, among which the University of Texas MD Anderson Cancer Center made the most contribution to the field with 122 papers. Lung Cancer published the highest number of SCLC-related articles with a total of 121, while the Journal of Thoracic Oncology received the most citations totaling 3,098. Rudin, Charles M., and Sage, Julien are the most creative author. Leora, Horn, 2018, New Engl J Med and Rudin, Charles M., Nat Genet, 2012 can be categorized as classic literature owing to their high citations or strong sigma value. “Heterogeneity & Subtypes” and “Immunotherapy” may be the new frontiers in the SCLC domain.ConclusionResearch on SCLC showed an upward trend based on the current global situation. Moreover, the current scope of collaboration in SCLC research is chiefly regional, which should further focus on transnational cooperation in the future. More attention should be devoted to “Heterogeneity & Subtypes” and “Immunotherapy”, which will be the hotspots in future research.

Monitoring animal behavior is crucial for protecting ecosystems, maintaining ecological balance, and improving animal welfare. By utilizing various monitoring devices, a wealth of behavioral data can be collected, which machine learning techniques can then analyze to identify specific behaviors. Artificial neural networks are particularly important in movement ecology. However, current research on animal behavior recognition faces several limitations. Many existing datasets are limited by homogeneous species categories, simplistic environmental conditions, restricted video perspectives, and a lack of alignment with the complexity of real-world environments. Consequently, there is significant room for improvement in the robustness and generalization of automatic animal behavior recognition models. To address these challenges, this paper introduces a diverse dataset of wild mammal behaviors. This dataset includes a wide variety of typical wild mammal species, providing a foundation for enhancing the generality and robustness of recognition models. The videos in this dataset capture different environmental contexts where wild mammals reside, across various times of day. Based on this dataset, a novel and highly efficient wild mammal behavior recognition model, EDNN, is proposed. The EDNN model integrates both temporal and spatial scales and achieves an average recognition accuracy of 79.17 % for basic locomotion behaviors, with a Top-1 accuracy of 81.37 % and a Top-5 accuracy of 98.04 %. These results demonstrate the feasibility of automating animal behavior recognition using large datasets collected from modern monitoring devices. The EDNN model is highly effective for behavior recognition and can be readily applied across diverse species and scenarios. It efficiently processes various video data and contributes to a deeper understanding of the movement ecology of species that are challenging to observe.

Abstract The absence of local features and global shape constraints severely limits the performance of the hourglass network for facial landmark detection in unconstrained environments. Moreover, diverse feature types and scales may result in low accuracy. This paper proposes a probability‐guided hourglass network to enhance the shape constraints for robust facial landmark detection. Firstly, a multi‐scale pre‐processing module is designed to extract features at different scales. Secondly, based on the heatmaps generated by the stacked hourglass network, the coarse localizations are obtained, while the probability maps are generated with local features. Finally, a probability‐based boundary regression method is proposed and the hausdorff distance is modified as the loss function to constrain the feature shape. Adaptive weights are also added to the loss function, which can help relieve the data imbalance problem. Subjective and objective experimental results on the challenging datasets show that this method outperforms the state‐of‐the‐arts on unconstrained conditions.

Traditional heatmap regression methods have some problems such as the lower limit of theoretical error and the lack of global constraints, which may lead to the collapse of the results in practical application. In this paper, we develop a facial landmark detection model aided by offset prediction to constrain the global shape. First, the hybrid detection model is used to roughly locate the initial coordinates predicted by the backbone network. At the same time, the head rotation attitude prediction module is added to the backbone network, and the Euler angle is used as the adaptive weight to modify the loss function so that the model has better robustness to the large pose image. Then, we introduce an offset prediction network. It uses the heatmap corresponding to the initial coordinates as an attention mask to fuze with the features, so the network can focus on the area around landmarks. This model shares the global features and regresses the offset relative to the real coordinates based on the initial coordinates to further enhance the continuity. In addition, we also add a multi-scale feature pre-extraction module to preprocess features so that we can increase feature scales and receptive fields. Experiments on several challenging public datasets show that our method gets better performance than the existing detection methods, confirming the effectiveness of our method.

The accurate prediction of potential associations between microRNAs (miRNAs) and small molecule (SM) drugs can enhance our knowledge of how SM cures endogenous miRNA-related diseases. Given that traditional methods for predicting SM-miRNA associations are time-consuming and arduous, a number of computational models have been proposed to anticipate the potential SM–miRNA associations. However, several of these strategies failed to eliminate noise from the known SM-miRNA association information or failed to prioritize the most significant known SM-miRNA associations. Therefore, we proposed a model of Graph Convolutional Network with Layer Attention mechanism for SM-MiRNA Association prediction (GCNLASMMA). Firstly, we obtained the new SM-miRNA associations by matrix decomposition. The new SM-miRNA associations, as well as the integrated SM similarity and miRNA similarity were subsequently incorporated into a heterogeneous network. Finally, a graph convolutional network with an attention mechanism was used to compute the reconstructed SM-miRNA association matrix. Furthermore, four types of cross validations and two types of case studies were performed to assess the performance of GCNLASMMA. In cross validation, global Leave-One-Out Cross Validation (LOOCV), miRNA-fixed LOOCV, SM-fixed LOOCV and 5-fold cross-validation achieved excellent performance. Numerous hypothesized associations in case studies were confirmed by experimental literatures. All of these results confirmed that GCNLASMMA is a trustworthy association inference method.

Abstract Although Sparse Representation based Classifier (SRC), a non‐parametric model, can obtain an interesting result for pattern recognition , a reasonable interpretation has been lacked for its classification mechanism. What is more, the training samples are used as off‐the‐shelf dictionary directly in SRC, which can make the feature hidden in the training samples hard be extracted. At the same time, the complexity of the algorithm is increased because of too many atoms of the dictionary. The authors first explains in detail the classification mechanism of SRC from the view of probabilistic collaborative subspace and offer the process to improve the stability of the algorithm using the joint probability in the case of the multi‐subspace. Then, the authors introduce the dictionary learning (DL) and Fisher criterion into the model to further enhance the discrimination of the coding coefficient. In order to ensure the convexity of the discrimination term and further enhance the discrimination, the authors add the L21‐norm term into the Fisher discrimination term and offer the proof for its convexity. Finally, the experimental result on a series of benchmark databases, such as AR, Extended Yale B, LFW3D‐hassner, LFW3D‐sdm and LFW3D‐Dlib, show that PCDDL outperforms existing classical classification models.

Abstract The traditional face alignment approaches based on cascade regression have achieved satisfactory result on the frontal face, but for the face with large changes in posture and expression, a single initial shape will lead to the result falling into local optimum. In order to solve this problem, a two‐stage cascade regression model for face alignment is proposed, which generates coarse initial shape from the aligned salient shape. The first stage is used to align the salient shape that contains some prominent landmarks. To enhance the robustness of authors' method, the fusion subspace is used to divide the samples, and each subset trains cascade regression model separately. The alignment results of the first stage are used to generate the coarse initial shapes for the second stage through 3D fitting. The second stage is still based on cascade regression, which is used to further predict the full shape. The experimental results demonstrate the proposed method can achieve state‐of‐art performance, especially in unconstrained conditions with various poses.

Defect detection on the surface of the steel strip is essential for the quality assurance of the steel strip. Precise localization and classification, the two significant tasks of defect detection, still need to be completed due to the diversity of defect scales. In this paper, a residual atrous spatial pyramid pooling (RASPP) module is first designed to enrich the multi-scale information of the feature maps and increase the receptive field of the feature maps. Secondly, a double pyramid network (DPN) that combines RASPP and feature pyramid is proposed to fuse multi-scale features further so that similar semantic features are shared among the features of each layer. Finally, DPN-Detector, an automatic surface defects detection network, is proposed, which embeds the DPN module into Faster R-CNN and replaces the original detection head with a designed double head. Experiments are carried out on the steel strip surface defect dataset (NEU-DET), and the results show that the mAP of DPN-Detector is as high as 80.93%, which is 3.52% higher than that of the baseline network Faster R-CNN. The classification accuracy is 74.64%, and the detection speed reaches 18.62 FPS. The proposed method performs better robustness, classification and regression capability than other steel strip defect detection methods.

Human action recognition plays a key role in human-computer interaction in complex environments. However, similar actions will lead to poor feature sequence extraction and result in a reduction in recognition accuracy. This paper proposes a method (Action-Fusion: Multi-label subspace Learning (MLSL)) from depth maps called Depth Sequential Information Entropy Maps (DSIEM) and skeleton data for human action recognition in multiple modal features. The DSIEM describe the spatial information of human motion with information entropy, and describe the temporal information through stitching. DSIEM can reduce the redundancy of depth sequences and effectively capture spatial motion states. MLSL studies the relationship between different modalities and the inherent connection between different labels. The method is evaluated on three public datasets: Microsoft action 3D dataset (MSR Action3D), University of Texas at Dallas-multimodal human action dataset (UTD-MHAD), UTD MHAD- Kinect Version-2 (UTD-MHAD-Kinect V2). Experimental results show that the proposed MLSL model obtains new state-of-the-art results, including achieving the average rate of the MSR Action3D to 93.55%, the average rate of the UTD-MHAD to 88.37% and the average rate of the UTD-MHAD-Kinect V2 to 90.66%.

In recent years, much research has been conducted on time series based human activity recognition (HAR) using wearable sensors. Most existing work for HAR is based on the manual labeling. However, the complete time serial signals not only contain different types of activities, but also include many transition and atypical ones. Thus, effectively filtering out these activities has become a significant problem. In this paper, a novel machine learning based segmentation scheme with a multi-probability threshold is proposed for HAR. Threshold segmentation (TS) and slope-area (SA) approaches are employed according to the characteristics of small fluctuation of static activity signals and typical peaks and troughs of periodic-like ones. In addition, a multi-label weighted probability (MLWP) model is proposed to estimate the probability of each activity. The HAR error can be significantly decreased, as the proposed model can solve the problem that the fixed window usually contains multiple kinds of activities, while the unknown activities can be accurately rejected to reduce their impacts. Compared with other existing schemes, computer simulation reveals that the proposed model maintains high performance using the UCI and PAMAP2 datasets. The average HAR accuracies are able to reach 97.71% and 95.93%, respectively.

Low-rank decomposition model has been widely used in fabric defect detection, where a matrix is decomposed into a low-rank matrix representing the defect-free region (background) of the image and a sparse matrix identifying the defective region (foreground). Two shortcomings, however, still exist. First, the traditional low-rank decomposition model can only deal with Laplacian noise, ignoring Gaussian noise in the image region. Second, when textiles have periodic complex patterns, it is difficult for traditional models to separate the defective regions from the background. To address these problems, we propose a low-rank decomposition model with noise regularization and gradient information (G-NLR): 1) noise regularization, which characterizes the image noise part and enlarges the gaps between defective objects and background in the feature space. 2) The gradient information constrains the noise term adaptively according to the mutation degree of the current pixel point, so as to guide the matrix decomposition and reduce the noise misjudgment. This ensures the accuracy of detection results. We evaluate our model for fabric defect detection on the standard database including box-, star- and dot-patterns, and compare with four recent detection methods. Experimental results show that our method has good results in terms of recall rate and precision.

With the rapid development of the computer and sensor field, inertial sensor data have been widely used in human activity recognition. At present, most relevant studies divide human activities into basic actions and transitional actions, in which basic actions are classified by unified features, while transitional actions usually use context information to determine the category. For the existing single method that cannot well realize human activity recognition, this paper proposes a human activity classification and recognition model based on smartphone inertial sensor data. The model fully considers the feature differences of different properties of actions, uses a fixed sliding window to segment the human activity data of inertial sensors with different attributes and, finally, extracts the features and recognizes them on different classifiers. The experimental results show that dynamic and transitional actions could obtain the best recognition performance on support vector machines, while static actions could obtain better classification effects on ensemble classifiers; as for feature selection, the frequency-domain feature used in dynamic action had a high recognition rate, up to 99.35%. When time-domain features were used for static and transitional actions, higher recognition rates were obtained, 98.40% and 91.98%, respectively.

In this paper, we propose a novel bilateral 2-D neighborhood preserving discriminant embedding for supervised linear dimensionality reduction for face recognition. It directly extracts discriminative face features from images based on graph embedding and Fisher's criterion. The proposed method is a manifold learning algorithm based on graph embedding criterion, which can effectively discover the underlying nonlinear face data structure. Both within-neighboring and between-neighboring information are taken into account to seek an optimal projection matrix by minimizing the intra-class scatter and maximizing the inter-class scatter based on Fisher's criterion. The performance of the proposed method is evaluated and compared with other face recognition schemes on the Yale, PICS, AR, and LFW databases. The experiment results demonstrate the effectiveness and superiority of the proposed method as compared with the state-ofthe-art dimensionality reduction algorithms.

Fabric defect detection plays a vital role in ensuring the production quality of the textile manufacturing industry. However, in practice, there are relatively few manually annotated defective samples, and considering both performance and parameter quantity, there is still room for optimization in the architecture of detection networks. Therefore, this paper proposes a textile defect detection method called MED-AGNet. Firstly, based on the diffusion model, a mask-embedding data augmentation method, MEDiffusion, is proposed. During the training process, a conditional term (M) that represents the shape of the defect is added, and through supervised learning, the generative model learns the correlation between the background and defects. In the generation stage, it samples from a normal distribution and relies on M guidance to gradually generate corresponding defective textiles, thereby expanding the original sample set. An attention-guided network (AGNet) is a network that utilizes attention to guide information across different scales. Its feature extraction module employs a dual-branch information residual unit (DIRU) as a substitute for the conventional convolution block, which combines the feature extraction capabilities of global pooling and max pooling, reducing the number of parameters while also achieving a certain improvement in detection results. In the feature fusion stage, it utilizes the attention-guided fusion module (AGFM), which can allow the attention information of high-level semantics to guide the low-level semantics, and simultaneously, adds a high-level semantic residual attention module (HSRA) to enhance the perception of defect shapes and improve detection effectiveness. Ultimately, AGNet’s true positive rate (TPR), positive predictive value (PPV), and f-measure exceed those of the state-of-the-art (SOTA) algorithms by 1.88%, 0.05%, and 0.77%, respectively, and with a consistent model architecture, its parameter quantity is reduced by 56%.

Aiming at the influence of uneven illumination on fabric feature extraction and the limitations of traditional frequency-based visual saliency algorithms, we propose a fabric defect detection method based on the combination of illumination correction and visual salient features—(1) Construct a multi-scale side window box (MS-BOX) filter to extract the illumination component of the image, then use the constructed two-dimensional gamma correction function to perform illumination correction on the image in the global angle, and finally enhance the local contrast of the image in the local angle; (2) Use the L0 gradient minimization method to remove the background texture of fabric images and highlight the defects; (3) Represent the fabric image as a quaternion image, where each pixel in the image is represented by a quaternion consisting of color, intensity and edge characteristics. The two-dimensional fractional Fourier transform (2D-FRFT) is used to obtain the saliency map of the quaternion image. Experiments show that our method has a higher overall recall rate for defect detection of star-patterned, box-patterned, and dot-patterned fabrics, and the overall recall-precision effect is better than other existing methods.

In contemporary research on human action recognition, most methods separately consider the movement features of each joint. However, they ignore that human action is a result of integrally cooperative movement of each joint. Regarding the problem, this paper proposes an action feature representation, called Motion Collaborative Spatio-Temporal Vector (MCSTV) and Motion Spatio-Temporal Map (MSTM). MCSTV comprehensively considers the integral and cooperative between the motion joints. MCSTV weighted accumulates limbs’ motion vector to form a new vector to account for the movement features of human action. To describe the action more comprehensively and accurately, we extract key motion energy by key information extraction based on inter-frame energy fluctuation, project the energy to three orthogonal axes and stitch them in temporal series to construct the MSTM. To combine the advantages of MSTM and MCSTV, we propose Multi-Target Subspace Learning (MTSL). MTSL projects MSTM and MCSTV into a common subspace and makes them complement each other. The results on MSR-Action3D and UTD-MHAD show that our method has higher recognition accuracy than most existing human action recognition algorithms.

This paper deals with the implementation of Steiner point of fuzzy set. Some definitions and properties of Steiner point are investigated and extended to fuzzy set. This paper focuses on establishing efficient methods to compute Steiner point of fuzzy set. Two strategies of computing Steiner point of fuzzy set are proposed. One is called linear combination of Steiner points computed by a series of crisp α-cut sets of the fuzzy set. The other is an approximate method, which is trying to find the optimal α-cut set approaching the fuzzy set. Stability analysis of Steiner point of fuzzy set is also studied. Some experiments on image processing are given, in which the two methods are applied for implementing Steiner point of fuzzy image, and both strategies show their own advantages in computing Steiner point of fuzzy set.

In this paper a synthetic model for stock prediction is proposed based on phase space reconstruction, Echo State Networks (ESN) and Moving Average Convergence/Divergence (MACD). In this model, time series data is reconstructed in phase space before feeding to the ESN. Guided by the MACD strategy, stock prices are predicted and valuable trading advices are provided. The proposed model is tested by Microsoft Company stock price. The experiments are compared with those provided by other models such as BackPropagation Neural Networks (BNN), Support Vector Machine (SVM), Hidden Markov Model (HMM) and Procedural Neural Networks (PNN). The test results show that the proposed model is effective and efficient in stock price prediction.