DRP-Net and clustering algorithm for Stem-Leaf segmentation and phenotypic trait extraction from tomato point clouds

A traffic surveillance system includes detection of vehicles which involves the detection and identification of license plate numbers. This paper proposes an intelligent approach of detecting vehicular number plates automatically using three efficient algorithms namely Ant colony optimization (ACO) used in plate localization for identifying the edges, a character segmentation and extraction algorithm and a hierarchical combined classification method based on inductive learning and SVM for individual character recognition. Initially the performance of the Ant Colony Optimization algorithm is compared with the existing algorithms for edge detection namely Canny, Prewitt, Roberts, Mexican Hat and Sobel operators. The Ant Colony Optimization used in communication systems has certain limitations when used in edge detection like random initial ant position in the image and the heuristic information being highly dictated by transition probabilities. In this paper, modifications like assigning a well-defined initial ant position and making use of weights to calculate heuristic value which will provide additional information about transition probabilities are used to overcome the limitations. Further a character extraction and segmentation algorithm which uses the concept of Kohonen neural network to identify the position and dimensions of characters is presented along with a comparison with the existing Histogram and Connected Pixels approach. Finally an inductive learning based classification method is compared with the Support Vector Machine based classification method and a combined classification method which uses both inductive learning and Support Vector Machine based approach for character recognition is proposed. The proposed character recognition algorithm may be more efficient than the other two.

The segmentation and extraction on color features can provide useful information for many different application domains. However, most of the existing image processing algorithms on feature extraction are gray image-based and consider only one-dimensional parameters. In order to carry out a fast and accurate color feature extraction, this paper proposes a color extraction algorithm by segmentation that is called a color extraction algorithm This algorithm is compared under different color distribution situations, and the extraction effect on color is also shown by the combination of the segmentation and feature extraction algorithms. Experimental results show that such segmentation algorithm has some advantages for color segmentation. In the fuzzy color image preprocessing, this paper gives the location method of region of interest. Moreover, compared with other existing extraction algorithms, the presented segmentation extraction algorithm in this paper not only has higher accuracy, shorter extraction time and stronger anti-interference ability, but also has better effect on more divergent color edge. Experimental evaluation of the proposed color extraction algorithm demonstrates dominance over existing algorithms for feature extraction. These researches in this paper provide a new way of thinking for color feature extraction by segmentation, which has an important theoretical references and practical significance.

Feature extraction and segmentation obviously play an utmost important role in autonomous mobile robot localization and navigation. In this paper we discuss some line segmentation and feature extraction algorithms and proposed an adaptive feature extraction algorithm for 2D laser range data. Features in indoor environments that are considered in this paper can be described as two geometric primitives: line segments which represent the walls, and corners. Segmentation process estimates the line segments belong to the walls and which represent the same object have been grouped together. Intersecting points of the line segments, called corners are treated as landmarks. Some segmentation algorithms are implemented and tested using laser range data captured in different environments and the effectiveness of them is analyzed. Finally, a better adaptive feature extraction and segmentation algorithm which generates a line map of an unstructured environment is proposed.

In order to increase the performance of document analysis systems a detailed quality evaluation of the achieved results is required. By focussing on segmentation algorithms, we point out that the results produced by the module under consideration should be evaluated directly; we will show that the text-based evaluation method which is often used in the document analysis domain does not accomplish the purpose of a detailed quality evaluation. Therefore, we propose a general evaluation approach for the comparison of segmentation results which is based on the segments directly. This approach is able to handle both algorithms that produce complete segmentations (partition) and algorithms that only extract objects of interest (extraction). Classes of errors are defined in a systematic way, and frequencies for each class can be computed. The evaluation approach is applicable to segmentation or extraction algorithms in a wide range. We have chosen the character segmentation task as an example in order to demonstrate the applicability of our evaluation approach, and we suggest to apply our approach to other segmentation tasks.

Summary form only given, as follows. Increasing the performance of document analysis systems requires a detailed quality evaluation of the achieved results. By focussing on segmentation algorithms, we point out that the results produced from the module under consideration should be evaluated directly; we show that the text based evaluation method which is often used in the document analysis domain is not sufficient for the purpose of a detailed quality evaluation of the segmentation module. Therefore, we propose a general evaluation approach for comparing segmentation results which is based on the segments directly. This approach is able to handle both algorithms which produce complete segmentations (partition) and algorithms which only extract objects of interest (extraction). Classes of errors are defined in a systematic way and frequencies for each class can be computed. The evaluation approach is applicable to segmentation or extraction algorithms in a wide range. We have chosen the character segmentation task as an example to demonstrate the applicability of our evaluation approach and we suggest applying our approach to other segmentation tasks.

It is crucial to evaluate human settlements to maintain human-based social development. To accurately extract human settlements land-cover information, this article focuses on the unsupervised image segmentation in high-resolution, remotely sensed imagery. J-based segmentation (JSEG) algorithm can offer good segmentation results, but it is highly dependent on image merging thresholds. In this article, a combined, unsupervised image segmentation algorithm is proposed, in which an initial segmentation result is first produced by JSEG and then the image texture is extracted for later region merging. The experiments on the high-resolution imagery show that, when compared with the results obtained by the original JSEG algorithm and eCognition, the proposed algorithm improves the segmentation accuracy.

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.

As a basic and important research, point cloud segmentation plays an important role in many fields. However, the traditional point cloud segmentation algorithm still suffers from two problems. The first problem is that voxel-based segmentation algorithms cannot guarantee accuracy in regions of inconsistent density. The other problem is the inefficiency of point-based clustering algorithms. Hence, in order to solve the first two problems, a new supervoxel-based segmentation algorithm is proposed. To address the first problem, a multi-resolution supervoxel algorithm is proposed to obtain the basic unit for clustering, which includes a new low-density region detection algorithm and a resegmentation process. However, over-detection during the construction of supervoxels leads to the presence of small fragments around large supervoxels. Therefore, for the second problem, a novel BPSO (belief propagation supervoxel optimization) algorithm is proposed to optimize the supervoxel. Moreover, an improved multi-resolution supervoxel- and graph-based segmentation (MGS) algorithm is presented for supervoxel clustering and a segmentation optimization algorithm is adopted to allocate the unallocated points. Experiments are conducted on different datasets, and segmentation results are evaluated quantitatively. Compared with traditional methods and advanced methods, the results show that this method can segment urban point clouds accurately and effectively.

Cardiovascular disease (CVD) is a common disease related to the heart and blood vessels. Due to the lack of research on the characteristics and effect of cardiovascular calcification on hemodialysis patients in China, it is almost impossible to accurately extract, segment, and measure the cardiovascular calcified regions from the cardiovascular calcification image. This article proposed an algorithm to extract and segment the calcified regions based on image characteristics. Firstly, the dome method was used to obtain the approximate region of calcification and the basic extraction and segmentation algorithm was used to preprocess the calcified region. Then, the region of calcification was enhanced using the image enhancement method before being further processed by the basic algorithm. After that, the preprocessed segmented image was compared back to the original image and only the initial grey value of the common area was kept in the original image. Finally, the basic segmentation algorithm was utilized again to process the original image before the threshold division and binarization being performed to obtain the final segmentation results. The results indicated that our method can segment the calcified region more accurately and thus more accurate distinguishment of the calcified regions from the non-calcified regions can be achieved.

Nowadays, TLS (terrestrial laser scanning) has been a relatively mature measuring equipment categorized to indoor measuring robots, but it is not widely adopted in indoor construction measurement at present. What accounts for its limited application are as follows: (1) the high cost of high‐accuracy laser LIDAR and (2) existing TLS equipment does not possess self‐adaptation scanning planning and takes no account of efficiency of point cloud processing and consumption of computing power. This paper proposes a novel TLS equipment and a high‐efficiency point cloud processing method customized for the novel equipment, with purpose to achieve self‐adaption measurement on the basis of indoor characteristics of construction during civil engineering at low cost. This paper mainly presents two parts of innovations: (1) for planning of scanning, the novel TLS features planning sampling density of scanning according to room size and converting scanning data from poses to point clouds, and (2) for processing of point clouds, this paper proposes two novel segmentation algorithms, namely, “on‐boundary segmentation algorithm” and “on‐angular‐distance segmentation algorithm,” based on indoor spatial structure features and characteristics of TLS. Besides, this paper presents modified RANSAC‐TLS (random sample consensus‐total least squares) plane fitting algorithm, on basis of TLS point cloud distribution characteristics and spatial transformation. Through actual measurement test, it is concluded that the “on‐boundary segmentation algorithm” and “on‐angular‐distance segmentation algorithm” are suitable for point cloud segmentation in different types of scenes. The modified RANSAC‐TLS have made a great improvement on accuracy of fitting versus LS (least squares), TLS (total least squares), and RANSAC‐LS. Finally, this paper conducts an experiment by executing an actual measurement and then preliminarily testifies the potential and future application of the proposed novel TLS (terrestrial laser scanning) equipment, with measurement parameters from it being changed in the experiment, by comparing with one existing TLS equipment—FARO. The test thus proves the relatively high feasibility and potential of the novel TLS presented in the paper (terrestrial laser scanning) in actual indoor measurement.

We present a fast algorithm for Korean text extraction and segmentation from subway signboards using smart phone sensors in order to minimize computational time and memory usage. The algorithm can be used as preprocessing steps for optical character recognition (OCR): binarization, text location, and segmentation. An image of a signboard captured by smart phone camera while holding smart phone by an arbitrary angle is rotated by the detected angle, as if the image was taken by holding a smart phone horizontally. Binarization is only performed once on the subset of connected components instead of the whole image area, resulting in a large reduction in computational time. Text location is guided by user's marker-line placed over the region of interest in binarized image via smart phone touch screen. Then, text segmentation utilizes the data of connected components received in the binarization step, and cuts the string into individual images for designated characters. The resulting data could be used as OCR input, hence solving the most difficult part of OCR on text area included in natural scene images. The experimental results showed that the binarization algorithm of our method is 3.5 and 3.7 times faster than Niblack and Sauvola adaptive-thresholding algorithms, respectively. In addition, our method achieved better quality than other methods.

Power line point cloud segmentation is one of the important tasks of airborne lidar (LiDAR) power patrol. In this paper, it is difficult to segment power line point clouds in the absence of power line point clouds, and the existing algorithm models are not comprehensive. A power line point cloud segmentation algorithm based on two models is proposed. Laser point cloud technology is widely used in transmission line operation and maintenance, and some power lines are often missing in laser point cloud data, which leads to the inaccuracy of traditional power line segmentation algorithms. In addition, at present, the academic circle mainly studies the non-coincident model of XOY plane projection of power line point cloud, but lacks the research on the coincident model of XOY plane projection. In this paper, two power line segmentation algorithms are proposed for two power line models (XOY plane projection is not coincident, XOY plane projection is coincident, hereinafter referred to as model 1 and model 2). Through practical engineering experiments and applications, the robustness and applicability of the proposed algorithm are verified, as well as the insensitivity to the lack of power line point clouds.

Threshold segmentation algorithm for automatic extraction of cerebral vessels from brain magnetic resonance angiography images.

The World Wide Web is becoming the most important source for surveying the state of research areas all over the world. DynamicView is such a Semantic Web portal implemented for researchers to query distribution of research areas in computer science. From millions of pages in free text available on thousands of Web sites, finding the distribution of research areas is a hard task. To resolve this problem, we proposed a page segmentation algorithm for information extraction, which is called DeSeA (delimiter based segmentation algorithm) as it divides a Web page into coherent blocks using pre-defined delimiters and finds relevant blocks from them. DeSeA carries out on an extended DOM tree (EDT), transforming it into a block tree. Compared with existing algorithms, it considers delimiter's content splitting level, with higher-level delimiters having higher priority to segment a Web page. Experiments show that DeSeA based information extraction algorithm fits commendably with DynamicView.

The high computational cost of the superpixel segmentation algorithms for hyperspectral remote sensing images makes them ideal candidates for parallel computation. The waterpixel algorithm, in particular, extracts segmentation regions called waterpixels and consists of four stages called vectorial gradient, spatial regularization, marker selection, and watershed transform. In this paper, an efficient version of a GPU algorithm for waterpixel segmentation using the Compute Unified Device Architecture (CUDA) is presented. The algorithm extracts all the spectral information available in the bands of the hyperspectral image through the vectorial gradient. A cellular automaton is selected for the computation of the watershed transform using a block-asynchronous implementation with 8-connectivity. The experimental analysis shows high speedup values for the resulting GPU algorithm when it is compared to a multicore OpenMP implementation using 8 threads.