Aiming at the problem of artificial artifacts due to phase disorder in frequency-domain speech enhancement algorithms， which limits the denoising performance and decreases the speech quality， a speech enhancement algorithm based on Multi-Scale Ladder-type Time-Frequency Conformer Generative Adversarial Network （MSLTF-CMGAN） was proposed. Taking the real part， imaginary part and magnitude spectrum of the speech spectrogram as input， the generator first learned the local and global feature dependencies between temporal and frequency domains by using time-frequency Conformer at multiple scales. Secondly， the Mask Decoder branch was used to learn the amplitude mask， and the Complex Decoder branch was directly used to learn the clean spectrogram， and the outputs of the two decoder branches were fused to obtain the reconstructed speech. Finally， the metric discriminator was used to judge the scores of speech evaluation metrics， and high-quality speech was generated by the generator through minimax training. Comparison experiments with various types of speech enhancement models were conducted on the public dataset VoiceBank+Demand by subjective evaluation Mean Opinion Score （MOS） and objective evaluation metrics.Experimental results show that compared with current state-of-the-art speech enhancement method CMGAN （Comformer-based MetricGAN）， MSLTF-CMGAN improves MOS prediction of the signal distortion （CSIG） and MOS predictor of intrusiveness of background noise （CBAK） by 0.04 and 0.07 respectively， even though its Perceptual Evaluation of Speech Quality （PESQ） and MOS prediction of the overall effect （COVL） are slightly lower than that of CMGAN， it still outperforms other comparison models in several subjective and objective speech evaluation metrics.

To solve the signal demodulation problem of asymmetric Paired Carry Multiple Access (PCMA) composed of the same frequency of main station and small station signals, a framework to realize this kind of signal demodulation was constructed. Parameter estimation is an indispensable part in the realization of two-way signal separation and demodulation for asymmetric PCMA communication systems. For the estimation accuracy of amplitude parameters, a searching amplitude estimation algorithm based on fourth-power method was proposed. Firstly, the demodulation model for asymmetric PCMA systems was established and the basic assumptions were made. Then the phase errors under different assumptions were compared with each other and the influence of phase error on the amplitude estimation algorithm was analyzed. Finally, a new amplitude estimation algorithm was proposed. Experimental results show that, under same Signal-to-Noise Ratio (SNR), the demodulation performance of the small station signal under normal phase error is inferior to its demodulation performance under mean value condition. When the order of magnitude of the Bit Error Rate (BER) is 10 ^-4, the demodulation performance of small station signal is improved by 1 dB with the improved algorithm, proving that the improved algorithm is better than fourth-power method.

In order to assist in the fast and efficient diagnosis of breast cancer and provide the prognosis information for pathologists, a computer-aided diagnosis approach for automatically grading breast pathological images was proposed. In the proposed algorithm,cells of pathological images were first automatically detected by deep convolutional neural network and sliding window. Then, the algorithms of color separation based on sparse non-negative matrix factorization, marker controlled watershed, and ellipse fitting were integrated to get the boundary of each cell. A total of 203-dimensional image-derived features, including architectural features of tumor, texture and shape features of epithelial cells were extracted from the pathological images based on the detected cells and the fitted boundary. A Support Vector Machine (SVM) classifier was trained by using the extracted features to realize the automated grading of pathological images. In order to verify the proposed algorithm, a total of 49 Hematoxylin & Eosin (H&E)-stained breast pathological images obtained from 17 patients were considered. The experimental results show that,for 100 ten-fold cross-validation trials, the features with the cell shape and the spatial structure of organization of pathological image set successfully distinguish test samples of low, intermediate and high grades with classification accuracy of 90.20%. Moreover, the proposed algorithm is able to distinguish high grade, intermediate grade, and low grade patients with accuracy of 92.87%, 82.88% and 93.61%, respectively. Compared with the methods only using texture feature or architectural feature, the proposed algorithm has a higher accuracy. The proposed algorithm can accurately distinguish the grade of tumor for pathological images and the difference of accuracy between grades is small.

The properties of the measured objects in 3D profile using the grating projection are more and more complex, there are a large number of splits in the extracted refinement grating stripes, and the refinement stripe encoding is very difficult. An automatic coding algorithm based on color structure light was proposed. The paper designed a new model of color structure light, introduced its design principle and implemented a new automatic stripe coding algorithm. First, the algorithm extracted the refinement grating stripe with color information from the color structure grating. Then, orderly encoded the refined stripes of each color by judging the best connected domain. Finally, the article got the stripe coding of the total image through combined coding by using the periodicity of grating model. The simulation experiment results show that the model design of color structure light is simple, the automatic coding algorithm of stripe has high accuracy and the error is decreased to 10 percent. The ideal 3D points cloud data model can be reconstructed through the strip coded data.

The cluster number is not generally set by K-means clustering algorithm beforehand, and artificial initial clustering number easily leads to the problem of unstable clustering results. A high-efficient algorithm for determining the K-means optimal clustering number was presented. The algorithm got the upper bound of the number of clustering search range through stratified sample data and designed a new kind of effective clustering indicator to evaluate the clustering degree of similarity between and within class after clustering. Thus the optimal number of clusters was obtained in the search range of the clusters number. The simulation results show that the algorithm can obtain the optimal clustering number fast and accurately, and the dataset clustering effect is good.

Most of the variants of Graph Cut algorithm do not impose any shape constraints on the segmentations, rendering it difficult to obtain semantic valid segmentation results. As for pedestrian segmentation, this difficulty leads to the non-human shape of the segmented object. An improved Graph Cut algorithm combining shape priors and discriminatively learned appearance model was proposed in this paper to segment pedestrians in static images. In this approach, a large number of real pedestrian silhouettes were used to encode the a'priori shape of pedestrians, and a hierarchical model of pedestrian template was built to reduce the matching time, which would hopefully bias the segmentation results to be humanlike. A discriminative appearance model of the pedestrian was also proposed in this paper to better distinguish persons from the background. The experimental results verify the improved performance of this approach.

Bad Mouthing and Self-Promoting (BS) collusion attack group and its detection mechanism, called BSCD, were proposed to resolve the security issues of the multiple malicious node collusion attack network nodes and affect their accurate positioning in the Wireless Sensor Network (WSN) reputation system. And the implementation method of the mechanism was given. It detected the abnormal recommended node, analyzed the evaluation behavior similarity between recommended nodes, and effectively detected the existence of collusion attack group, thereby reduced its damage and impact on the reputation of the system. The simulation results show that, BSCD has significant effect on the detection and resisting BS collusion attack group, effectively improves the malicious node detection rate in the reputation system and the capacity of the entire system to resist malicious node.

An approximate outage probability formulation and diversity order were derived for multi-hop diversity system employing Decode-and-Forward (DF) relaying in high Signal-to-Noise Ratio (SNR). And a modified multi-hop Selective Decode-and-Forward (SDF) protocol was proposed to overcome the limitation of space full diversity system unemploying DF relaying, the source node resended the information when the relay could not decode information correctly. An approximate outage probability formulation and diversity order were deduced for SDF protocol in high SNR. The theoretical analysis and simulation results show that the SDF protocol can improve transmission performance and achieve the full space diversity contrasted with DF protocol.

A new low-power localization algorithm based on the evaluation of distributed reputation was proposed to improve the security and energy consumption of the node positioning for wireless sensor network. The concepts of Trustworthy Node Table (TNT) and the backup cluster head node were introduced to find the reliable beacon nodes quickly, and the backup cluster head node could assist and monitor the cluster head node, reducing the workload of the cluster head and participating in the integration process of the beacon nodes reputation values. The proposed algorithm enhanced the reliability and integrity of the beacon nodes, improved the efficiency and security of the node localization, reduced the systems energy consumption and improved the detection rate of malicious nodes. The simulation results show that in malicious node environment, the algorithm can effectively improve the detection rate of malicious nodes, reduce the positioning error, weaken the malicious nodes damage and influence on the positioning system to achieve the safe positioning of the nodes.

Feature extraction and similarity measurement for graphs are important issues in computer vision and pattern recognition. However, traditional methods could not describe the graphs under some non-rigid transformation adequately, so a new graph feature descriptor and its similarity measurement method were proposed based on Graph Context (GC) descriptor. Firstly, a sample point set was obtained by discretely sampling. Secondly, graph context descriptor was presented based on the sample point set. At last, improved Earth Mover's Distance (EMD) was used to measure the similarity for graph context descriptor. Different from the graph edit distance methods, the proposed method did not need to define cost function which was difficult to set in those methods. The experimental results demonstrate that the proposed method performs better for the graphs under some non-rigid transformation.

The traditional feature selection algorithms are limited to single-label data. Concerning this problem, multi-label ReliefF algorithm was proposed for multi-label feature selection. For multi-label data, based on label co-occurrence, this algorithm assumed the label contribution value was equal. Combined with three new methods calculating the label contribution, the updating formula of feature weights was improved. Finally a distinguishable feature subset was selected from original features. Classification experiments demonstrate that, with the same number of features, classification accuracy of the proposed algorithm is obviously higher than the traditional approaches.

A method of recognizing agricultural products image based on multi-instance learning is proposed for solving problems with which agricultural products selling in supermarket encounter. An improved Single Blob with Neighbors (SBN) method was adopted to organize bags and meanwhile extract features of an image. The target concept was learned by maximizing Diverse Density(DD) and applied to images’ recognition. Experiments were performed on both multi-class produce image dataset by self-collection and single-class produce image selected from Amsterdam Library of Object Image (ALOI). The experiments show that, the method is able to recognize multi-class produce images captured under various illumination conditions and distracters-scattered background. Compared with global method, the method can attain a higher recognition rate of 95.45%. The results indicate that recognition of produce image based on Multiple Instance Learning aims for aiding automatic sale in supermarket is feasible.