Real-time combustion state identification via image processing: A dynamic data-driven approach

The purpose of this article is to explore effective image feature extraction algorithms in the context of big data, and to mine their potential information from complex image data. Based on the BRISK and SIFT algorithms, this paper proposes an image feature extraction and matching algorithm based on BRISK corner points. By combining the SIFT scale space and the BRISK algorithm, a new scale space construction method is proposed. The BRISK algorithm extracts the corner invariant features. Then, by using the improved feature matching method and eliminating the mismatching algorithm, the exact matching of the images is realized. A large number of experimental verifications were performed in the standard test Mikolajczyk image database and aerial image database. The experimental results show that the improved algorithm in this paper is an effective image matching algorithm. The highest accuracy of actual aerial image matching can reach 85.19%, and it can realize the actual aerial image matching that BRISK and SIFT algorithms cannot complete. The improved algorithm in this paper has the advantages of higher matching accuracy and strong robustness.

Cephalometric landmark identification on cephalometric analysis Downs has been developed. Here we identify a 10 point landmarks used in cephalometric Downs analysis: Gnation, Gonion, Menton, Nation, Orbital, Porion, Point A, Point B, Pogonion, and Sella. The research was conducted in three phases: preprocessing, feature extraction and identification. In the preprocessing phase, the things we do are: image size normalization, Enhance contrast, and took the ROI for each landmark. We use of PPED algorithms for image feature extraction and to measure the similarity between the template vector and vector test, euclidean distance gives good results, despite the large number of image samples do not always give the results was always good. To improve systems performance, we use a multithreading technique. The experimental results showed that the methods used can work well although still found deficiencies and inaccuracies. Accuracy of projections mainly on bilateral landmarks greatly affect the outcome of the identification.

Intelligent transportation is a system that combines data-driven information with traffic management to achieve intelligent monitoring and retrieval functions. In order to further improve the retrieval accuracy of the system model, a new retrieval model was designed. The functional requirements of the system were summarized, and the three stages of data preprocessing, feature matching, and feature extraction were analyzed in detail. The study adopted preprocessing measures such as equalization and normalization to minimize the negative effects of noise and brightness. Based on the performance of various algorithms, the distance method was selected as the feature matching method, which has a wider applicability and is better at processing bulk data. Next, the study utilizes Euclidean distance method to extract keyframes and divides the feature extraction into three parts: color, shape, and texture. The methods of color moment, canny operator, and grayscale co-occurrence matrix are used to extract them, and ultimately achieve relevant image retrieval. The research conducted multiple experiments on the retrieval performance of the model, and analyzed the results of retrieving single and mixed features. The experimental results showed that the algorithm performed better in the face of mixed feature extraction. Compared with the average value of a single feature, the recall and precision of the three mixed features increased by 13.78% and 15.64%, respectively. Moreover, in the case of a large number of concurrent features, the algorithm also met the basic requirements. When the concurrent number was 100, the average response time of the algorithm is 4.46 seconds. Therefore, the algorithm proposed by the research institute effectively improves the ability of video retrieval and can meet the requirements of timeliness, which can be widely applied in practical applications.

Breast imaging technique called mammography has gained bigger attention among the researchers for the diagnosis of breast malignancy in the woman. Mammogram screening is the most effective procedure to visualize various potential problems in the breast. The two most common features connected with breast tumors are mass lesions and microcalcification. The collection of suitable image preprocessing, segmentation, feature extraction, selection and prediction algorithms play an essential role in the accurate detection and classification of cancer on mammograms. Classification techniques estimate unlabeled datasets class labeling depending on its similarity to the pattern learned. The Glowworm Swarm Optimization(GSO) algorithm is ideal for finding several solutions, and dissimilar or equivalent objective function values at the same time. This feature of GSO is useful for optimizing the feature set obtained from multiscale feature extraction procedures. Poor performance in generalization is the issue that arises due to the unconditioned output matrix of the hidden stage of the ELM classifier. The optimization algorithms will address this matter because of their global search capabilities. This article suggests ELM with the Fruitfly Optimization Algorithm (ELM‐FOA) along with GSO to regulate the input weight to achieve maximal performance at the hidden node of the ELM. The testing precision and sensitivity of GSO‐ELM‐FOA are 100% and 97.91%, respectively. The system developed will detect the calcifications and tumors with an accuracy of 99.15%.

Breast cancer has been identified as one of the major diseases that have led to the death of women in recent decades. Mammograms are extensively used by physicians to diagnose breast cancer. The selection of appropriate image enhancement, segmentation, feature extraction, feature selection and prediction algorithm plays an essential role in precise cancer diagnosis on mammograms and remains as a major task in the research field. Classification methods predict the class label for unlabeled dataset based on its proximity to the learnt pattern. The selected features obtained after feature selection are classified using an extreme learning machines (ELM) to three classes with the classes being normal, benign and malignant. Low generalisation performance is the problem which happens due to the ill-conditioned output matrix of the hidden layer of the classifier. The optimisation algorithms would resolve these issues because of their global searching ability. This paper proposes ELM with Fruitfly Optimisation Algorithm (ELM-FOA) to tune the input weight to obtain optimum output at the ELM’s hidden node to obtain the solution analytically. The testing sensitivity and precision of ELM-FOA are 97.5% and 100% respectively. The developed method can detect the calcifications and tumours with 99.04% accuracy. The optimal selection of preprocessing and segmentation algorithms, features from multiple feature filters and the efficient classifier algorithm meliorate the performance of the approach.

We present a two-stage mapping model (TSMM), which is intended to minimise the semantic gap for content-based image retrieval (CBIR) by reducing recognition errors during the image indexing stage. This model is composed of a feature extraction module based on our image segmentation and feature extraction algorithm, a colour and texture classification modules based on support vector machines (SVMs), and an inference module based on fuzzy logic to make final decisions as high level concepts from the colour and texture concepts. The experimental results show that the proposed method outperforms general approaches by using one single SVM classifier as direct mapping between the combined colour and texture feature vectors and high level concepts directly.

During the past few years, advancements in high end computers and sensor techniques made it possible to develop a real-time odontological biometric identification and verification system apart from the existing offline forensic odontological systems. However, this requires highly automated teeth image segmentation and feature extraction algorithms. In this paper, we propose a novel non-forensic biometric technique, employing 3D optical sensors for data acquisition and representing it as 3D surface meshes (stereo lithography format). Initially, feature regions are estimated using geometric curvature information and principal component analysis. Subsequently, active contours optimize these feature regions to segment exactly each tooth surface. Finally, a feature vector is computed for the set of teeth which is derived from the relative position and tilt orientation of each tooth relative to a well defined reference system.

Image recognition is a feature extraction and pattern matching technique to classify various objectives in computer science. In theory, it mainly relies on the images with distinguishable features as a good starting point. Each image has its own unique characteristics, and image features can be categorized into color features, texture features, and shape features etc. Therefore, using modern recognition techniques, we extract the image features through various algorithms to search for images with high similarities. There are many image recognition algorithms, and some of them are with high recognition rate while others are robust. But not all algorithms can be unconditionally implemented without adjusting to the real situation. In this paper, we introduce the classification algorithm based on Support Vector Machine (SVM) and the feature extraction method based on Principal Components Analysis (PCA). We employ the feature extraction algorithm to characterize facial features and recognize faces by comparing them to those stored in the training data set. Finally, we show the applications of feature extraction and classification algorithms in character recognition. The real characters printed on a cord are preprocessed by PCA, and then classified and identified by SVM with a good recognition rate if properly processed.

This thesis aims to develop an embedded image processing board, which can be used to realize various image feature extraction and processing algorithms for robotic applications. The circuitry of this image board can be divided into two parts, one is for a digital signal processor (DSP) and the other for a field programmable gate array (FPGA). A low power, small size CMOS image sensor with YUV4:2:2 image format and 512x480 pixel size is adopted for image acquisition. The DSP module is mainly used for implementing image processing algorithms, while the FPGA module for lower level image processing and sensor control functions. In this work, the TMS320C6414TGLZ7 DSP from Texas Instruments Corporation and the EP2C35F672C6 FPGA from Altera Corporation were selected for the image board. In this work, speeded-up robust features (SURF) algorithm has been successfully realized on the developed image board. An USB interface with average throughput of 22MB/sec works to transfer acquired image frames and extracted SURF descriptors to a personal computer, where feature matching is executed. Experimental results show that the refresh period of frame grabbing is about 33.5ms. When the number of SURF descriptor is equal to 64, the computing time is about 417.5ms; when the number of SURF descriptor is equal to 199, the computing time is about 985ms. The typical matching rate varies from 70.2% to 20%, depending on the threshold of the patterns used in feature extraction.

Local Feature Extraction (LFE) algorithm is an effective feature extraction method developed in recent years. One of the shortcomings of the current LFE algorithm is that it can only process labeled data, and does not work well when the amount of the labeled data is limited. However, it is usually easy to obtain large amount of unlabeled data but only a few labeled data. In this paper, we propose a new feature extraction algorithm, called Semi-Supervised LFE (SSLFE), which can handle both labeled and unlabeled data to perform feature extraction. In the proposed algorithm, the labeled data are used to maximize the margin and the unlabeled data are used as regulations with respect to the intrinsic geometric structure of the data. The final projection matrix can be obtained by eigenvalue decomposition. Experiments on several datasets demonstrate that SSLFE achieves much higher classification accuracy than LFE.

This paper first analyzes the use and expression of line and surface, color, and text as graphic design composition elements. Secondly, it analyzes the relationship between interaction design and other disciplines and constructs a dynamic interactive design model for graphic design. Then, the features of color, texture, shape and spatial relationship of the image are extracted by the feature extraction algorithm, and the gesture recognition algorithm is used to determine the operation requested by the user and execute it. Finally, data analysis was conducted on 32 randomly selected students from the design school of University H who applied the dynamic interactive design model. The results show that the dynamic interactive design has the highest rating in terms of smooth operation, with an average score of 6.893, and the lowest rating is in terms of low learning cost, with an average score of 5.087. This shows that the dynamic interactive design model of graphic elements can improve the smoothness of operation, thus improving the dynamic interactive experience of students and providing direction for the development of the field of graphic design.

Preserving cultural heritage and historic sites is an important problem. These sites are subject to erosion and vandalism, and, as long-lived artifacts, they have gone through many phases of construction, damage and repair. We believe that it is important to use 3D model building technology to create an accurate record of these sites, so preservationists can track changes and foresee structural problems. From a digital libraries perspective, 3D models also allow a much wider audience to virtually see and tour these sites. We have developed a suite of new methods that can reduce the time to build a model through automation. Our methods utilize range image segmentation and feature extraction algorithms. Once the 3D model is constructed, it is necessary to texture map it with imagery to create a geometrically and photometrically correct model. We are developing a 3D visualization environment to aid archaeologists in their post-excavation interpretation and analysis.

Mining knowledge for HEp-2 cell image classification

A complete mine detection/classification (D/C) system has been specified and implemented, which runs in real-time, and has been exercised on the latest available dual-frequency side-scan sonar acoustic image sets. The compete DC system is comprised of a collection of algorithms that has been developed and evolved at Draper Laboratory over the past decade. The detection process consists of image normalization, enhancement, segmentation, and feature extraction algorithms. The enhancement algorithm is a variant of a Markov Random Field based anomaly screener developed in FY-94. The feature that were extracted were those derived in FY-93. A distance constrained matching algorithm, which was developed in FY-95, is used to generate a list of high and low frequency fused tokens. The classification process involves the evaluation of a hierarchy of three multi-layer perceptron neural networks: HF, LF, and HF/LF fused. Research performed in FY-95 also concentrated on the development of several variants of information fusion with hierarchical neural networks. The 'discriminant-combining' variant of fusion was selected as part of this DC system. In addition, a classification post- processing and decision node statistic modification step, which was developed in FY-96, was included. This paper will describe the algorithm that were implemented. However, the emphasis will be on the performance results of processing the latest available side-scan imagery, comparison of single sensor vs dual-frequency sensor results, and the issues that were encountered while exercising the DC system on the new data set.

This paper presents an image feature extraction algorithm for robotic applications. The proposed method robust against scene illumination change, viewpoint change, specularity, colour saturation, imperfect focus of image, and shadows. The proposed algorithm is a bottom up approach which consists of three phases: colour classification, correction, and description. In colour classification phase, chrominance (CIE Lab colour space) is used to segment coloured images in order to detect potential coloured region. Thresholding and several morphological operations are applied in correction phase in order to eliminate the noise pixels. Finally, moment method is used to identify the desired image features (area, position and orientation) of targets. The results are presented to illustrate each operations involved and demonstrate the performance of proposed image feature extraction algorithm. Overall, the proposed method is suitable for known operating environment cases where a robot operates in predefined workspace with prior knowledge of target.