Semi-supervised Label Propagation Research Articles

Semi-supervised label propagation for multi-source remote sensing image change detection

Remote sensing image change detection remains a challenging task. Most existing approaches are based on fully supervised learning, but labeled data are so scarce for change detection. It is difficult to exhibit high detection performance with a limited amount of labeled data. In this paper, we propose a semi-supervised Label Propagation (SSLP) approach for multi-source remote sensing image change detection. First, a clustering label propagation (CLP) method is designed to cluster pre and post images, respectively, and assign pseudo labels to unlabeled pixel pairs that have similar mapping relationships to labeled pixel pairs. Second, a pixel density metric is investigated to filter out the data with low density and retain the data with high density, which can ensure the reliability of the propagated data. Third, a secondary expansion method based on pixel neighborhood is used to generate enough training data for training a classifier. Finally, the effectiveness of SSLP is validated on three real datasets by comparing to other related methods.

Joint auto-weighted graph fusion and scalable semi-supervised learning

Graph carries out a key role in graph-based semi-supervised label propagation, as it clarifies the structure of the data manifold. The performance of label propagation methods depends on the adopted graph and can be enhanced by merging different graphs that are obtained from multiple sources of information. While there exist algorithms that perform graph fusion they have several weaknesses. Most of these algorithms define graph fusion and label propagation as two separate tasks. Moreover, when the number of data expands, these strategies are not well-suited due to the use of transductive learning in the label propagation phase which makes the label prediction for unseen samples difficult. Furthermore, very few algorithms extract the information contained in the label space. Additionally, most of the graph fusion techniques adopt equal or static weights for different views, which is not the best choice as distinctive features (hence different graphs) contain various information. To overcome these shortcomings, we propose an Auto-weighted Multi-view Semi-Supervised Learning method (AMSSL), which is based on an inductive learning algorithm (i.e., Flexible Manifold Embedding) and profited a projection matrix for predicting the labels of out-of-sample data. The proposed AMSSL method represents a unified framework that dynamically fuses various information obtained from different features and also from the label space and adaptively designates appropriate weights according to the usefulness of each view. Our experimental results on seven small and large image datasets demonstrate the superiority of the proposed method compared to the use of one single feature and other state-of-the-art graph fusion methods.

Improved well-log classification using semisupervised label propagation and self-training, with comparisons to popular supervised algorithms

Machine-learning techniques allow geoscientists to extract meaningful information from data in an automated fashion, and they are also an efficient alternative to traditional manual interpretation methods. Many geophysical problems have an abundance of unlabeled data and a paucity of labeled data, and the lithology classification of wireline data reflects this situation. Training supervised algorithms on small labeled data sets can lead to overtraining, and subsequent predictions for the numerous unlabeled data may be unstable. However, semisupervised algorithms are designed for classification problems with limited amounts of labeled data, and they are theoretically able to achieve better accuracies than supervised algorithms in these situations. We explore this hypothesis by applying two semisupervised techniques, label propagation (LP) and self-training, to a well-log data set and compare their performance to three popular supervised algorithms. LP is an established method, but our self-training method is a unique adaptation of existing implementations. The well-log data were made public through an SEG competition held in 2016. We simulate a semisupervised scenario with these data by assuming that only one of the 10 wells has labels (i.e., core samples), and our objective is to predict the labels for the remaining nine wells. We generate results from these data in two stages. The first stage is applying all the algorithms in question to the data as is (i.e., the global data), and the results from this motivate the second stage, which is applying all algorithms to the data when they are decomposed into two separate data sets. Overall, our findings suggest that LP does not outperform the supervised methods, but our self-training method coupled with LP can outperform the supervised methods by a notable margin if the assumptions of LP are met.

Remote sensing image change detection using a hybrid graphical model

Semisupervised graph learning has a broad prospect in remote sensing (RS) image change detection. However, an improper graph model may result in a contradiction between the detection accuracy and computational efficiency. In order to effectively extract the structural information of changes and heavily reduce the computational burden, we propose a hybrid graphical model (HGM) for bitemporal RS image change detection. The HGM utilizes the hybrid superpixels (HSPs) as its vertices, and a hybrid graph kernel (HGK) function is proposed for measuring the similarities between the vertices. The HSPs are composed of the background superpixels and foreground isolated pixels of a subtraction image. The HGM effectively exploits the image structures, and the small graph size dramatically reduces the computational complexity. Moreover, the piecewise HGK function well detects the structures of the changed areas and heavily resists the background disturbances. A semisupervised label propagation algorithm is implemented with the HGK matrix for obtaining the final change detection results. Experimental results on real RS images demonstrate the effectiveness and efficiency of the proposed method and prove that it is a good candidate for RS image change detection.

Network-based direction of movement prediction in financial markets

Market prediction has been an important research problem for decades. Having better predictive models that are both more accurate and faster has been attractive for both researchers and traders. Among many approaches, semi-supervised graph-based prediction has been used as a solution in recent researches. Based on this approach, we present two prediction models. In the first model, a new network structure is introduced that can capture more information about markets’ direction of movements compared to the previous state of the art methods. Based on this novel network, a new algorithm for semi-supervised label propagation is designed that is able to prediction the direction of movement faster and more accurately.The second model is a mixture of experts system that decides between supervised or semi-supervised approaches. Besides this, the model gives us the ability to identify the markets that their data are helpful in constructing the network. Our models are shown to be both faster regarding computational complexity and running time and more accurate in prediction comparing to best rival models in literature of graph-based semi-supervised prediction. The results are also tested to be statistically significant.

Multi Similarity Metric Fusion in Graph-Based Semi-Supervised Learning

In semi-supervised label propagation (LP), the data manifold is approximated by a graph, which is considered as a similarity metric. Graph estimation is a crucial task, as it affects the further processes applied on the graph (e.g., LP, classification). As our knowledge of data is limited, a single approximation cannot easily find the appropriate graph, so in line with this, multiple graphs are constructed. Recently, multi-metric fusion techniques have been used to construct more accurate graphs which better represent the data manifold and, hence, improve the performance of LP. However, most of these algorithms disregard use of the information of label space in the LP process. In this article, we propose a new multi-metric graph-fusion method, based on the Flexible Manifold Embedding algorithm. Our proposed method represents a unified framework that merges two phases: graph fusion and LP. Based on one available view, different simple graphs were efficiently generated and used as input to our proposed fusion approach. Moreover, our method incorporated the label space information as a new form of graph, namely the Correlation Graph, with other similarity graphs. Furthermore, it updated the correlation graph to find a better representation of the data manifold. Our experimental results on four face datasets in face recognition demonstrated the superiority of the proposed method compared to other state-of-the-art algorithms.

Integrating Semi-supervised and Supervised Learning Methods for Label Fusion in Multi-Atlas Based Image Segmentation.

A novel label fusion method for multi-atlas based image segmentation method is developed by integrating semi-supervised and supervised machine learning techniques. Particularly, our method is developed in a pattern recognition based multi-atlas label fusion framework. We build random forests classification models for each image voxel to be segmented based on its corresponding image patches of atlas images that have been registered to the image to be segmented. The voxelwise random forests classification models are then applied to the image to be segmented to obtain a probabilistic segmentation map. Finally, a semi-supervised label propagation method is adapted to refine the probabilistic segmentation map by propagating its reliable voxelwise segmentation labels, taking into consideration consistency of local and global image appearance of the image to be segmented. The proposed method has been evaluated for segmenting the hippocampus in MR images and compared with alternative machine learning based multi-atlas based image segmentation methods. The experiment results have demonstrated that our method could obtain competitive segmentation performance (average Dice index > 0.88), compared with alternative multi-atlas based image segmentation methods under comparison. Source codes of the methods under comparison are publicly available at www.nitrc.org/frs/?group_id=1242.

INTEGRATING SEMI-SUPERVISED LABEL PROPAGATION AND RANDOM FORESTS FOR MULTI-ATLAS BASED HIPPOCAMPUS SEGMENTATION.

A novel multi-atlas based image segmentation method is proposed by integrating a semi-supervised label propagation method and a supervised random forests method in a pattern recognition based label fusion framework. The semi-supervised label propagation method takes into consideration local and global image appearance of images to be segmented and segments the images by propagating reliable segmentation results obtained by the supervised random forests method. Particularly, the random forests method is used to train a regression model based on image patches of atlas images for each voxel of the images to be segmented. The regression model is used to obtain reliable segmentation results to guide the label propagation for the segmentation. The proposed method has been compared with state-of-the-art multi-atlas based image segmentation methods for segmenting the hippocampus in MR images. The experiment results have demonstrated that our method obtained superior segmentation performance.

Text/Figure Separation in Document Images Using Docstrum Descriptor and Two-Level Clustering

We propose a novel algorithm for text/figure separation tailored for binary document images containing line drawings, block diagrams, charts, schemes and other kinds of business graphics. Most of the approaches for this task rely either on clever design of visual descriptor allowing to easily distinguish text and graphics regions or on the supervised learning using dataset of labeled text/figure regions. Such approaches often provide moderate separation accuracy when applied to document images which contain very diverse set of figure classes and lack sufficiently representative labeled training dataset. In contrast, our method is well-suited for vast variety of figure classes and capable of operating either in semi-supervised mode or unsupervised mode. We achieve this by leveraging unsupervised learning algorithms applied to Docstrum descriptors extracted from regions of interest and subsequent semi-supervised label propagation or unsupervised label inference. Another advantage of our method is its suitability for large scale data processing which is achieved through efficient kernel-approximating feature mapping applied to Docstrum descriptors and two-level clustering where fast mini-batch K-means algorithm is first applied to large scale data and only small number of resulting cluster centroids is subsequently processed by one of the more sophisticated clustering algorithms.

Detecting fuzzy network communities based on semi-supervised label propagation

Most existing methods use only network topology information, which neglect much other important information such as the network background information which could be useful to uncover the network communities. In this paper, we proposed a fast semi-supervised algorithm to uncover the fuzzy network c ommunities using label propagation technology, which incorporating the prior information to facilitate the community detection process. Specially, we know the true community membership of a certain percentage nodes in advance. Firstly, the node graph is transformed to line graph, and the weight of line graph are defined. Then, the proposed algorithm is applied in the line graph to propagate the label of certain labeled nodes to the whole network until convergence, and the final label of a given node is its community id. It worth mentioning that our algorithm is very fast and with almost linear time complexity. Extensive simulations using both synthetic and real-world benchmark networks are performed to verify the algorithmic performance.

Acquiring Knowledge of Affective Events from Blogs Using Label Propagation

Many common events in our daily life affect us in positive and negative ways. For example, going on vacation is typically an enjoyable event, while being rushed to the hospital is an undesirable event. In narrative stories and personal conversations, recognizing that some events have a strong affective polarity is essential to understand the discourse and the emotional states of the affected people. However, current NLP systems mainly depend on sentiment analysis tools, which fail to recognize many events that are implicitly affective based on human knowledge about the event itself and cultural norms. Our goal is to automatically acquire knowledge of stereotypically positive and negative events from personal blogs. Our research creates an event context graph from a large collection of blog posts and uses a sentiment classifier and semi-supervised label propagation algorithm to discover affective events. We explore several graph configurations that propagate affective polarity across edges using local context, discourse proximity, and event-event co-occurrence. We then harvest highly affective events from the graph and evaluate the agreement of the polarities with human judgements.

Automatic Segmentation of Hippocampus for Longitudinal Infant Brain MR Image Sequence by Spatial-Temporal Hypergraph Learning.

Accurate segmentation of infant hippocampus from Magnetic Resonance (MR) images is one of the key steps for the investigation of early brain development and neurological disorders. Since the manual delineation of anatomical structures is time-consuming and irreproducible, a number of automatic segmentation methods have been proposed, such as multi-atlas patch-based label fusion methods. However, the hippocampus during the first year of life undergoes dynamic appearance, tissue contrast and structural changes, which pose substantial challenges to the existing label fusion methods. In addition, most of the existing label fusion methods generally segment target images at each time-point independently, which is likely to result in inconsistent hippocampus segmentation results along different time-points. In this paper, we treat a longitudinal image sequence as a whole, and propose a spatial-temporal hypergraph based model to jointly segment infant hippocampi from all time-points. Specifically, in building the spatial-temporal hypergraph, (1) the atlas-to-target relationship and (2) the spatial/temporal neighborhood information within the target image sequence are encoded as two categories of hyperedges. Then, the infant hippocampus segmentation from the whole image sequence is formulated as a semi-supervised label propagation model using the proposed hypergraph. We evaluate our method in segmenting infant hippocampi from T1-weighted brain MR images acquired at the age of 2 weeks, 3 months, 6 months, 9 months, and 12 months. Experimental results demonstrate that, by leveraging spatial-temporal information, our method achieves better performance in both segmentation accuracy and consistency over the state-of-the-art multi-atlas label fusion methods.

Label propagation based on collaborative representation for face recognition

Recently, collaborative representation (CR) has been shown to produce impressive performance on face recognition. However, the performances of CR depend on the number of labeled training samples for each class. When the labeled training samples per class are insufficient, CR would perform inaccurately and correspondingly degrades the final recognition performance. To solve this problem, in this paper, we introduce the CR into semi-supervised learning and propose a novel semi-supervised label propagation approach based on collaborative representation. Based on the subspace assumption that samples of the same class lie in the same subspace, each labeled sample can be well represented by the unlabeled samples of the same class. Our algorithm exploits a large amount of unlabeled samples which contain much more useful information as a dictionary to represent labeled samples, and propagates the label information from labeled data to unlabeled data. Thus, the information of unlabeled data can be effectively explored in our method, which can further improve the performance of collaborative representation with limited labeled training samples. Furthermore, we introduce our label propagation into other semi-supervised learning algorithm to further improve its, recognition performance. Experimental results are presented to demonstrate the efficacy of the proposed method.

Using composite low rank and sparse graph for label propagation

Based on the low rank representation (LRR) and the sparse representation (SR), a composite LRR with SR graph LRRSR for semi-supervised label propagation is proposed. The LRRSR aims to capture both the global structure of the data by a low rank constraint and the local structure of the data by a sparse constraint simultaneously. A composite framework is applied to fuse the two graphs. Then, a label propagation framework is used to transmit the labels from the labelled samples to the unlabelled samples. It is applied on several face image datasets and the experimental results demonstrate its good performance for face classification with a limited number of labelled samples.

Image Retrieval Based on Semi-Supervised Orthogonal Discriminant Embedding

An retrieval algorithm based on dimensionality reduction is proposed to effectively extract the features to improve the performance of image retrieval. Firstly, the most important properties of the subspaces with respect to image retrieval is captured by intelligently utilizing the similarity and dissimilarity information of semantic and geometric structure in image database. Secondly, We propose Semi-supervised Orthogonal Discriminant Embedding Label Propagation method (SODELP) for image retrieval. The experimental results show that our method has the discrimination power against colour, texture and shape features and has good retrieval performance.

Robust image annotation via simultaneous feature and sample outlier pursuit

Graph-based semi-supervised image annotation has achieved great success in a variety of studies, yet it essentially and intuitively suffers from both the irrelevant/noisy features (referred to as feature outliers) and the unusual/corrupted samples (referred to as sample outliers). In this work, we investigate how to derive robust sample affinity matrix via simultaneous feature and sample outlier pursuit. This task is formulated as a Dual-outlier and Prior-driven Low-Rank Representation (DP-LRR) problem, which possesses convexity in objective function. In DP-LRR, the clean data are assumed to be self-reconstructible with low-rank coefficient matrix as in LRR; while the error matrix is decomposed as the sum of a row-wise sparse matrix and a column-wise sparse matrix, the ℓ 2,1 -norm minimization of which encourages the pursuit of feature and sample outliers respectively. The DP-LRR is further regularized by the priors from side information, that is, the inhomogeneous data pairs. An efficient iterative procedure based on linearized alternating direction method is presented to solve the DP-LRR problem, with closed-form solutions within each iteration. The derived low-rank reconstruction coefficient matrix is then fed into any graph based semi-supervised label propagation algorithm for image annotation, and as a by-product, the cleaned data from DP-LRR can also be utilized as a better image representation to generally boost image annotation performance. Extensive experiments on MIRFlickr, Corel30K, NUS-WIDE-LITE and NUS-WIDE databases well demonstrate the effectiveness of the proposed formulation for robust image annotation.

Prediction of Alzheimer's diagnosis using semi-supervised distance metric learning with label propagation

Alzheimer's disease (AD) is the most common form of dementia and leads to irreversible neurogenerative damage of the brain. However, the current diagnostic tools have poor sensitivity, especially for the early stages of AD and do not allow for diagnosis until AD has lead to irreversible brain damage. Therefore, it is crucial that AD is detected as early as possible. Although it is very hard, laborious and time-consuming to gather many AD and non-AD labeled samples, gathering unlabeled samples is easier than labeled samples. Since standard learning algorithms learn a diagnosis model from labeled samples only, they require many labeled samples and do not work well when the number of training samples is small. Therefore, it is very desirable to develop a predictive learning method to achieve high performance using both labeled samples and unlabeled samples. To address these problems, we propose semi-supervised distance metric learning using Random Forests with label propagation (SRF-LP) which incorporates labeled data for obtaining good metrics and propagates labels based on them. Experimental results showed that SRF-LP outperformed standard supervised learning algorithms, i.e., RF, SVM, Adaboost and CART and reached 93.1% accuracy at a maximum. Especially, SRF-LP largely outperformed when the number of training samples is very small. Our results also suggested that SRF-LP exhibits a synergistic effect of semi-supervised distance metric learning and label propagation.