Classifier ensemble based-on bi-coded chromosome genetic algorithm for automatic image annotation

Image distance metric learning based on neighborhood sets for automatic image annotation

Dynamic Adaboost learning with feature selection based on parallel genetic algorithm for image annotation

Automating the process of annotation of images is a crucial step towards efficient and effective management of increasingly high volume of content. A graph-based approach for automatic image annotation is proposed which models both feature similarities and semantic relations in a single graph. The proposed approach models the relationship between the images and words by an undirected graph. Semantic information is extracted from paired nodes. The quality of annotation is enhanced by introducing graph link weighting techniques. The proposed method achieves fast solution by using incremental fast random walk with restart (IFRWR) algorithm, without apparently affecting the accuracy of image annotation.

Due to the prevalence of digital cameras, it is easy to retrieve digital images from the Internet. With the rapid development of digital image processing, databases, and Internet technologies, how to efficiently manage a large amount of digital images is very important. In this paper, we proposed a novel approach for automatic image annotation. We extract color, texture, and shape features from a set of training images to build the main object classifier and background object models by using Support Vector Machine (SVM). We apply JSEG to segment background objects out of images, and then extract the feature vectors from the segmented objects for identification. In order to prevent over-segmenting the main object, the combination of Active Contour Model and JSEG is proposed to improve the system performance. Since the images in the same class have background consistency, we exploit Gaussian mixture model (GMM) to explore the relationship between image classes and image backgrounds, and build the association knowledge base. After classifying test images, we only need to compare the backgrounds with the related models for classification. Finally, the experimental results show that the proposed method has high effectiveness for image annotation.

This article presents an automatic image level annotation approach that takes advantage of both context and semantics presented in segmented images. The proposed approach is based on the optimization of classes’ scores using particle swarm optimization. In addition, random forest classifier and normalized cuts algorithm have been applied for automatic image classification, annotation, and clustering. For the proposed approach, each input image is segmented using the normalized cuts segmentation algorithm in order to create a descriptor for each segment. Two parameter selection models have been selected for particle swarm optimization algorithm and many voting techniques have been implemented to find the most suitable set of annotation words per image. Experimental results, using Corel5k benchmark annotated images dataset, demonstrate that applying optimization algorithms along with random forest classifier achieved noticeable increase in image annotation performance measures compared to related researches on the same dataset.

Image annotation systems aim at automatically annotating images with some predefined keywords. In this paper, we propose an automatic image annotation approach by incorporating word correlations into multi-class support vector machine (SVM). At first, each image is segmented into five fixed-size blocks or tiles and MPEG-7 visual descriptors are applied to represent color and texture features of blocks. Keywords are manually assigned to every block of training images. Then, multi-class SVM classifier is trained for semantic concepts. Word or concept correlations are computed by a co-occurrence matrix. The probability outputs from SVM and word correlations are combined to obtain the final results. The minimal-redundancy-maximum-relevance (mRMR) method is used to reduce feature dimensions. The experiments on Corel 5000 dataset demonstrate our approach is effective and efficient.

Automatic image annotation is one of the research fields helping to extract the meaning of images, which aims at the production of a set of semantic annotations for an image to help better present the concept. Over the past few decades, researchers have developed a number of approaches for automatic image annotation. Nevertheless, previous studies have not fully accounted for visual features and annotated features. Therefore, it is still possible to achieve a better annotation performance by combining visual and annotated information. In this study, our aim is to associate multiple semantic tags with a given image. In particular, we detect how to obtain the image annotation by utilizing visual and annotated information. To take advantage of visual information, we first designed a modified neural network method to acquire the features of the image content. In addition, to obtain the annotated features, we exploit an aggregated network embedding approach that consists of annotation embedding, social embedding, profile embedding, and semantic embedding. Finally, to produce an accurate image annotation, we integrate the two aforementioned methods, that is, combining the visual and annotated information, to build a unified cooperative training framework. The experimental results on three real-world datasets clarify that our presented method is superior to the current popular image annotation approaches.

ISABoost: A weak classifier inner structure adjusting based AdaBoost algorithm—ISABoost based application in scene categorization

Power system security is a major concern in real‐time operation. It is essential to protect the system from blackout by taking proper control actions. This study proposes a boosting algorithm for the precise and accurate prediction of static security assessment of power systems using synchronised measurements. In addition to security status, the proposed approach also predicts the type of violations which may be either line overload/voltage violation or both of the insecure operating conditions. To overcome the computational complexity, the number of input phasor measurements is reduced by a statistical approach based on class separability and correlation coefficient indices. In the classification stage, support vector machines (SVMs) are used as weak classifiers and a strong classifier is constructed as the linear combination of many weak SVM classifiers. The performance of the Adaptive Boosting (AdaBoost) algorithm is further improved by a new weight updation strategy using fuzzy clustering thresholding technique. The efficiency of the proposed approach is demonstrated on IEEE 14‐bus, IEEE 30‐bus, and Indian 246‐bus systems. Further, the test results reveal that the proposed method of security assessment performs better than the other traditional classifiers viz. SVM, feed forward neural network and k‐nearest neighbour classifier.

A novel random fast multi-label deep forest classification algorithm

This paper presents a novel multi-label classification framework for domains with large numbers of labels. Automatic image annotation is such a domain, as the available semantic concepts are typically hundreds. The proposed framework comprises an initial clustering phase that breaks the original training set into several disjoint clusters of data. It then trains a multi-label classifier from the data of each cluster. Given a new test instance, the framework first finds the nearest cluster and then applies the corresponding model. Empirical results using two clustering algorithms, four multi-label classification algorithms and three image annotation data sets suggest that the proposed approach can improve the performance and reduce the training time of standard multi-label classification algorithms, particularly in the case of large number of labels.

Alzheimer's disease (AD), a psychiatric problem, availing between those who are 65 and older. Additionally, the disease's steady development of a variety of visible and invisible symptoms, such as irritability and aggression, has a substantial negative impact on a patient's overall quality of life. Although many treatments have been developed to help reduce its symptoms, AD has no known cure. As a result, the field of AD management is growing, and a comprehensive framework for the early detection of AD must be created. In this study, we created three classification models for predicting AD using machine learning and five models for predicting AD using ensemble learning. SVM, DTs, and RF are the three basis classifiers employed in the current work. Five ensemble classifiers, XGBoost, Voting Classifier, Extra Trees (ETs) Classifier, Gradient Boost, and AdaBoost, are then thoroughly compared. After thoroughly inspecting the dataset for outliers or other noise, a feature selection method known as PCA and several preprocessing techniques are used to lessen the issue of overfitting and performance enhancement. Additionally, this study utilised the longitudinal information from the OASIS website, which included 150 patients overall, 72 of whom were not demented and 78 of whom were. The RF model, which had an accuracy of 83.92% compared to the other two base classifiers, provided the best classification performance, while the ETs Classifier, an ensemble classifier, performed the best when compared to base and ensemble classifiers, with an accuracy of 86.60%.

The paper presents a new machine learning method to solve the pose estimation problem. The method is based on the soft margin AdaBoost (SMA) algorithm (Ratsch, G. et al., Machine Learning, vol.42, no.3, p.287-320, 2001). The AdaBoost algorithm has been used with great success as a high-level learning procedure to obtain strong classifiers from weak classifiers, but it tends to overfit in the presence of very noisy data. Recent studies show that a regularised AdaBoost algorithm, such as SMA, can achieve better results for noisy data. We propose two new techniques for classifying the image as frontal (face is within /spl plusmn/25/spl deg/) or profile; one is based on the original Adaboost algorithm, the other on SMA. It is shown that the SMA based technique is more robust than the one based on the original AdaBoost, and yields better results. All the techniques were trained and tested on four databases. Experimental results show that the classification error of the SMA method is less than 2% for suitable parameters, regardless of the conditions associated with the face. In addition, the method performs extremely well even when some facial features become partially or wholly occluded.

When weak classifiers, i.e., estimators, are not justifying the classification of breast cancer then ensemble learning is a way to improve the classification of cancer. The ensemble is basically an aggregator where all weak classifiers are merged to get a strong classifier. The ensemble is based on a majority voting scheme. A hard voting scheme is used to take a major vote of each classifier whereas a soft voting scheme takes the weights of the probability of each classifier. A custom based weights are assigned in this paper and the final classification of cancer using ensemble classifier is outperformed than each estimator. The highest accuracy from the proposed ensemble classifier is achieved up to 96.47% where the lowest estimator got 93.18 %. The AUC score of ensemble classifier achieved is 0.9948 which is one of the highest among all other estimators.

Multi-label image classification aims to classify all categories in images simultaneously. When current multi-label classification methods meet fine-grained objects in a single image, the extreme inter-class similarity and over-prediction problems are two major challenges that hinder model performance. To solve the above two problems, we propose Voronoi density based Locally Unique Network (VoLUNet). First, due to high correlation between predictions of different classes, following the Kolmogorov-Arnold Network (KAN), we design the Weak Inter-class Correlation Classifier (WIC-Classifier) to replace linear weights setting in MLP architecture, promoting the potential of fine-grained discrimination. Second, we propose a Local Non-Maximum Suppression (Local-NMS) loss to multi-label classification model, predicting only one unique class with high prediction value for each local region. Third, different classes may have different pixel proportions and Local-NMS loss will be imbalanced for diverse fine-grained classes, we design the Voronoi Density based Superpixel Module (VDSM) to balance the quantities of local feature vectors with different classes. Finally, comprehensive experiments are conducted on four datasets, TreeSatAI, GeoLifeCLEF, FothemNet ShipRSImageNet, and our VoLUNet can significantly improve the classification performance compared to current state-of-the-art models. Codes of this paper are public available at https://github.com/cv516Buaa/BinghaoLiu/tree/main/VoLUNet.