Fast video shot boundary detection framework employing pre-processing techniques

Video shot boundary detection is the initial and fundamental step towards video indexing, browsing and retrieval. Great efforts have been paid on developing accurate shot boundary detection algorithms. However, the high computational cost in shot detection becomes a bottleneck for real-time applications. The problem of making a balance between detection accuracy and speed is addressed in this paper, and a novel fast detection framework is presented. The general framework that employs pre-processing techniques can improve both detection speed and precision. In the pre-processing stage, adaptive local thresholding is adopted to classify non-boundary segments and candidate segments that may contain shot boundaries. The candidate segments are refined using bisection-based comparisons to eliminate non-boundary frames. Only refined candidate segments are preserved for further detections; hence, the speed of shot detection is improved by reducing detection scope. Moreover, prior knowledge about each possible shot boundary such as its type and duration can be obtained in the pre-processing stage, which can accelerate the consequent hard cut and gradual transition detections. Experimental results indicate that the proposed framework is effective in accelerating the shot detection process, and it can also achieve excellent detection accuracies.     

Robust structure from motion with affine camera via low-rank matrix recovery

We present a novel approach to structure from motion that can deal with missing data and outliers with an affine camera. We model the corruptions as sparse error. Therefore the structure from motion problem is reduced to the problem of recovering a low-rank matrix from corrupted observations. We first decompose the matrix of trajectories of features into low-rank and sparse components by nuclear-norm and l1-norm minimization, and then obtain the motion and structure from the low-rank components by the classical factorization method. Unlike pervious methods, which have some drawbacks such as depending on the initial value selection and being sensitive to the large magnitude errors, our method uses a convex optimization technique that is guaranteed to recover the low-rank matrix from highly corrupted and incomplete observations. Experimental results demonstrate that the proposed approach is more efficient and robust to large-scale outliers.     

Evolutionary-computer-assisted design of image operators that detect interest points using genetic programming

This work describes a way of designing interest point detectors using an evolutionary-computer-assisted design approach. Nowadays, feature extraction is performed through the paradigm of interest point detection due to its simplicity and robustness for practical applications such as: image matching and view-based object recognition. Genetic programming is used as the core functionality of the proposed human-computer framework that significantly augments the scope of interest point design through a computer assisted learning process. Indeed, genetic programming has produced numerous interest point operators, many with unique or unorthodox designs. The analysis of those best detectors gives us an advantage to achieve a new level of creative design that improves the perspective for human-machine innovation. In particular, we present two novel interest point detectors produced through the analysis of multiple solutions that were obtained through single and multi-objective searches. Experimental results using a well-known testbed are provided to illustrate the performance of the operators and hence the effectiveness of the proposal.     

Bayesian learning of inverted Dirichlet mixtures for SVM kernels generation

We describe approaches for positive data modeling and classification using both finite inverted Dirichlet mixture models and support vector machines (SVMs). Inverted Dirichlet mixture models are used to tackle an outstanding challenge in SVMs namely the generation of accurate kernels. The kernels generation approaches, grounded on ideas from information theory that we consider, allow the incorporation of data structure and its structural constraints. Inverted Dirichlet mixture models are learned within a principled Bayesian framework using both Gibbs sampler and Metropolis-Hastings for parameter estimation and Bayes factor for model selection (i.e., determining the number of mixture’s components). Our Bayesian learning approach uses priors, which we derive by showing that the inverted Dirichlet distribution belongs to the family of exponential distributions, over the model parameters, and then combines these priors with information from the data to build posterior distributions. We illustrate the merits and the effectiveness of the proposed method with two real-world challenging applications namely object detection and visual scenes analysis and classification.     

Discriminative feature co-occurrence selection for object detection

This paper describes an object detection framework that learns the discriminative co-occurrence of multiple features. Feature co-occurrences are automatically found by Sequential Forward Selection at each stage of the boosting process. The selected feature co-occurrences are capable of extracting structural similarities of target objects leading to better performance. The proposed method is a generalization of the framework proposed by Viola and Jones, where each weak classifier depends only on a single feature. Experimental results obtained using four object detectors, for finding faces and three different hand gestures, respectively, show that detectors trained with the proposed algorithm yield consistently higher detection rates than those based on their framework while using the same number of features.     

HMM-based writer identification in music score documents without staff-line removal

Writer identification from musical score documents is a challenging task due to its inherent problem of overlapping of musical symbols with staff-lines. Most of the existing works in the literature of writer identification in musical score documents were performed after a pre-processing stage of staff-lines removal. In this paper we propose a novel writer identification framework in musical score documents without removing staff-lines from the documents. In our approach, Hidden Markov Model (HMM) has been used to model the writing style of the writers without removing staff-lines. The sliding window features are extracted from musical score-lines and they are used to build writer specific HMM models. Given a query musical sheet, writer specific confidence for each musical line is returned by each writer specific model using a log-likelihood score. Next, a log-likelihood score in page level is computed by weighted combination of these scores from the corresponding line images of the page. A novel Factor Analysis-based feature selection technique is applied in sliding window features to reduce the noise appearing from staff-lines which proves efficiency in writer identification performance. In our framework we have also proposed a novel score-line detection approach in musical sheet using HMM. The experiment has been performed in CVC-MUSCIMA data set and the results obtained show that the proposed approach is efficient for score-line detection and writer identification without removing staff-lines. To get the idea of computation time of our method, detail analysis of execution time is also provided.     

A multi-layer approach to opinion polarity classification using augmented semantic tree kernels

Tremendous increase in user-generated content (UGC) published over the web in the form of natural language has posed a formidable challenge to automated information extraction (IE) and content analysis (CA). Techniques based on tree kernels (TK) have been successfully used for modelling semantic compositionality in many natural language processing (NLP) applications. Essentially, these techniques obtain the similarity of two production rules based on exact string comparison between the peer nodes. However, semantically identical tree fragments are forbidden even though they can contribute to the similarity of two trees. A mechanism needs to be addressed that accounts for the similarity of rules with varied syntax and vocabulary holding knowledge that are relatively analogous. In this paper, a hierarchical framework based on document object model (DOM) tree and linguistic kernels that jointly address subjectivity detection, opinion extraction and polarity classification is addressed. The model proceeds in three stages: during first stage, the contents of each DOM tree node is analysed to estimate the complexity of vocabulary and syntax using readability test. In second stage, the semantic tree kernels extended with word embeddings are used to classify nodes containing subjective and objective content. Finally, the content returned to be subjective is further examined for opinion polarity classification using fine-grained linguistic kernels. The efficiency of the proposed model is demonstrated through a series of experiments being conducted. The results reveal that the proposed polarity-enriched tree kernel (PETK) results in better prediction performance compared to the conventional tree kernels.     

Multiresolution feature detection using a family of isotropicbandpass filters

Presents a novel method which uses a special type of multi-scale isotropic band-pass filters to detect image landmark and corner features. This paper mainly contributes to the solutions of the following problems: (1) defining a general family of feature detectors under a unified framework that are able to enhance and detect the desired features; (2) explaining theoretically and experimentally why these feature points can be detected by the proposed detectors; and (3) extending the detectors to multi-scale versions for jointly achieving good detectability and localization where an automatic scale selection method is applied. The paper then presents several applications for detecting landmarks and corners using the proposed methods, in order to illustrate their usage. These include detecting landmarks from gesture images (of the face and hand), from airborne and vehicle-borne IR landmine images, and from images containing object corners. Experiments have been performed using the proposed detectors to these applications. Some comparisons and evaluations have also been performed. The results have demonstrated the effectiveness of the proposed detectors in terms of feature detectability and localization.     

Object detection using feature subset selection

Past work on object detection has emphasized the issues of feature extraction and classification, however, relatively less attention has been given to the critical issue of feature selection. The main trend in feature extraction has been representing the data in a lower dimensional space, for example, using principal component analysis (PCA). Without using an effective scheme to select an appropriate set of features in this space, however, these methods rely mostly on powerful classification algorithms to deal with redundant and irrelevant features. In this paper, we argue that feature selection is an important problem in object detection and demonstrate that genetic algorithms (GAs) provide a simple, general, and powerful framework for selecting good subsets of features, leading to improved detection rates. As a case study, we have considered PCA for feature extraction and support vector machines (SVMs) for classification. The goal is searching the PCA space using GAs to select a subset of eigenvectors encoding important information about the target concept of interest. This is in contrast to traditional methods selecting some percentage of the top eigenvectors to represent the target concept, independently of the classification task. We have tested the proposed framework on two challenging applications: vehicle detection and face detection. Our experimental results illustrate significant performance improvements in both cases.