Semi-supervised learning based probabilistic latent semantic analysis for automatic image annotation

In recent years,multimedia annotation problem has been attracting significant research attention in multimedia and computer vision areas,especially for automatic image annotation,whose purpose is to provide an efficient and effective searching environment for users to query their images more easily. In this paper,a semi-supervised learning based probabilistic latent semantic analysis( PLSA) model for automatic image annotation is presenred. Since it's often hard to obtain or create labeled images in large quantities while unlabeled ones are easier to collect,a transductive support vector machine( TSVM) is exploited to enhance the quality of the training image data. Then,different image features with different magnitudes will result in different performance for automatic image annotation. To this end,a Gaussian normalization method is utilized to normalize different features extracted from effective image regions segmented by the normalized cuts algorithm so as to reserve the intrinsic content of images as complete as possible. Finally,a PLSA model with asymmetric modalities is constructed based on the expectation maximization( EM) algorithm to predict a candidate set of annotations with confidence scores. Extensive experiments on the general-purpose Corel5k dataset demonstrate that the proposed model can significantly improve performance of traditional PLSA for the task of automatic image annotation.

Online belief propagation algorithm for probabilistic latent semantic analysis

Probabilistic latent semantic analysis （PLSA） is a topic model for text documents, which has been widely used in text mining, computer vision, computational biology and so on. For batch PLSA inference algorithms, the required memory size grows linearly with the data size, and handling massive data streams is very difficult. To process big data streams, we propose an online belief propagation （OBP） algorithm based on the improved factor graph representation for PLSA. The factor graph of PLSA facilitates the classic belief propagation （BP） algorithm. Furthermore, OBP splits the data stream into a set of small segments, and uses the estimated parameters of previous segments to calculate the gradient descent of the current segment. Because OBP removes each segment from memory after processing, it is memoryefficient for big data streams. We examine the performance of OBP on four document data sets, and demonstrate that OBP is competitive in both speed and accuracy for online ex- pectation maximization （OEM） in PLSA, and can also give a more accurate topic evolution. Experiments on massive data streams from Baidu further confirm the effectiveness of the OBP algorithm.

User-oriented web search based on PLSA

In order to solve the problem that current search engines provide query-oriented searches rather than user-oriented ones, and that this improper orientation leads to the search engines＇ inability to meet the personalized requirements of users, a novel method based on probabilistic latent semantic analysis （PLSA） is proposed to convert query-oriented web search to user-oriented web search. First, a user profile represented as a user＇ s topics of interest vector is created by analyzing the user＇ s click through data based on PLSA, then the user＇ s queries are mapped into categories based on the user＇ s preferences, and finally the result list is re-ranked according to the user＇ s interests based on the new proposed method named user-oriented PageRank （UOPR）. Experiments on real life datasets show that the user-oriented search system that adopts PLSA takes considerable consideration of user preferences and better satisfies a user＇ s personalized information needs.

Complex human activities recognition using interval temporal syntactic model

A novel method based on interval temporal syntactic model was proposed to recognize human activities in video flow. The method is composed of two parts: feature extract and activities recognition. Trajectory shape descriptor, speeded up robust features(SURF) and histograms of optical flow(HOF) were proposed to represent human activities, which provide more exhaustive information to describe human activities on shape, structure and motion. In the process of recognition, a probabilistic latent semantic analysis model(PLSA) was used to recognize sample activities at the first step. Then, an interval temporal syntactic model, which combines the syntactic model with the interval algebra to model the temporal dependencies of activities explicitly, was introduced to recognize the complex activities with a time relationship. Experiments results show the effectiveness of the proposed method in comparison with other state-of-the-art methods on the public databases for the recognition of complex activities.

Fusing PLSA model and Markov random fields for automatic image annotation

A novel image auto-annotation method is presented based on probabilistic latent semantic analysis(PLSA) model and multiple Markov random fields(MRF).A PLSA model with asymmetric modalities is first constructed to estimate the joint probability between images and semantic concepts,then a subgraph is extracted served as the corresponding structure of Markov random fields and inference over it is performed by the iterative conditional modes so as to capture the final annotation for the image.The novelty of our method mainly lies in two aspects:exploiting PLSA to estimate the joint probability between images and semantic concepts as well as multiple MRF to further explore the semantic context among keywords for accurate image annotation.To demonstrate the effectiveness of this approach,an experiment on the Corel5 k dataset is conducted and its results are compared favorably with the current state-of-the-art approaches.

Beyond bag of latent topics: spatial pyramid matching for scene category recognition

We propose a heterogeneous, mid-level feature based method for recognizing natural scene categories. The proposed feature introduces spatial information among the latent topics by means of spatial pyramid, while the latent topics are obtained by using probabilistic latent semantic analysis （pLSA） based on the bag-of-words representation. The proposed feature always performs better than standard pLSA because the performance of pLSA is adversely affected in many cases due to the loss of spatial information. By combining various interest point detectors and local region descriptors used in the bag-of-words model, the proposed feature can make further improvement for diverse scene category recognition tasks. We also propose a two-stage framework for multi-class classification. In the first stage, for each of possible detector/descriptor pairs, adaptive boosting classifiers are employed to select the most discriminative topics and further compute posterior probabilities of an unknown image from those selected topics. The second stage uses the prod-max rule to combine information coming from multiple sources and assigns the unknown image to the scene category with the highest ＇final＇ posterior probability. Experimental results on three benchmark scene datasets show that the proposed method exceeds most state-of-the-art methods.

Re-examining urban region and inferring regional function based on spatial-temporal interaction

Urban system is shaped by the interactions between different regions and regions planned by the government,then reshaped by human activities and residents’needs.Understanding the changes of regional structure and dynamics of city function based on the residents’movement demand are important to evaluate and adjust the planning and management of urban services and internal structures.This paper constructed a probabilistic factor model on the basis of probabilistic latent semantic analysis and tensor decomposition,for purpose of understanding the higher order interactive population mobility and its impact on urban structure changes.First,a four-dimensional tensor of time(T)×week(W)×origin(O)×destination(D)was constructed to identify the day-to-day activities in three time modes and weekly regularity of weekday/weekend pattern.Then we reclassified the urban regions based on the space clustering formed by the space factor matrix and core tensor.Finally,we further analysed the space–time interaction on different time scales to deduce the actual function and connection strength of each region.Our research shows that the application of individual-based spatial–temporal data in human mobility and space–time interaction study can help to analyse urban spatial structure and understand the actual regional function from a new perspective.