Convergence and tracking analysis of a variable normalised LMF (XE-NLMF) algorithm

The least-mean-fourth (LMF) algorithm is known for its fast convergence and lower steady state error, especially in sub-Gaussian noise environments. Recent work on normalised versions of the LMF algorithm has further enhanced its stability and performance in both Gaussian and sub-Gaussian noise environments. For example, the recently developed normalised LMF (XE-NLMF) algorithm is normalised by the mixed signal and error powers, and weighted by a fixed mixed-power parameter. Unfortunately, this algorithm depends on the selection of this mixing parameter. In this work, a time-varying mixed-power parameter technique is introduced to overcome this dependency. A convergence analysis, transient analysis, and steady-state behaviour of the proposed algorithm are derived and verified through simulations. An enhancement in performance is obtained through the use of this technique in two different scenarios. Moreover, the tracking analysis of the proposed algorithm is carried out in the presence of two sources of nonstationarities: (1) carrier frequency offset between transmitter and receiver and (2) random variations in the environment. Close agreement between analysis and simulation results is obtained. The results show that, unlike in the stationary case, the steady-state excess mean-square error is not a monotonically increasing function of the step size. (c) 2007 Elsevier B.V. All rights reserved.

Influence of the Graphical Levels of Detail of a Virtual Thrower on the Perception of the Movement

Virtual reality has a number of advantages for analyzing sports interactions such as the standardization of experimental conditions, stereoscopic vision, and complete control of animated humanoid movement. Nevertheless, in order to be useful for sports applications, accurate perception of simulated movement in the virtual sports environment is essential. This perception depends on parameters of the synthetic character such as the number of degrees of freedom of its skeleton or the levels of detail (LOD) of its graphical representation. This study focuses on the influence of this latter parameter on the perception of the movement. In order to evaluate it, this study analyzes the judgments of immersed handball goalkeepers that play against a graphically modified virtual thrower. Five graphical representations of the throwing action were defined: a textured reference level (L0), a nontextured level (L1), a wire-frame level (L2), a moving point light display (MLD) level with a normal-sized ball (L3), and a MLD level where the ball is represented by a point of light (L4). The results show that judgments made by goalkeepers in the L4 condition are significantly less accurate than in all the other conditions (p

The exponentiated convex variable step-size (ECVSS) algorithm

For some time there is a large interest in variable step-size methods for adaptive filtering. Recently, a few stochastic gradient algorithms have been proposed, which are based on cost functions that have exponential dependence on the chosen error. However, we have experienced that the cost function based on exponential of the squared error does not always satisfactorily converge. In this paper we modify this cost function in order to improve the convergence of exponentiated cost function and the novel ECVSS (exponentiated convex variable step-size) stochastic gradient algorithm is obtained. The proposed technique has attractive properties in both stationary and abrupt-change situations. (C) 2010 Elsevier B.V. All rights reserved.

On-line Emotion Recognition in a 3-D Activation-Valence-Time Continuum using Acoustic and Linguistic Cues

For many applications of emotion recognition, such as virtual agents, the system must select responses while the user is speaking. This requires reliable on-line recognition of the user’s affect. However most emotion recognition systems are based on turnwise processing. We present a novel approach to on-line emotion recognition from speech using Long Short-Term Memory Recurrent Neural Networks. Emotion is recognised frame-wise in a two-dimensional valence-activation continuum. In contrast to current state-of-the-art approaches, recognition is performed on low-level signal frames, similar to those used for speech recognition. No statistical functionals are applied to low-level feature contours. Framing at a higher level is therefore unnecessary and regression outputs can be produced in real-time for every low-level input frame. We also investigate the benefits of including linguistic features on the signal frame level obtained by a keyword spotter.

Face Recognition in Global Harmonic Subspace

In this paper, a novel pattern recognition scheme, global harmonic subspace analysis (GHSA), is developed for face recognition. In the proposed scheme, global harmonic features are extracted at the semantic scale to capture the 2-D semantic spatial structures of a face image. Laplacian Eigenmap is applied to discriminate faces in their global harmonic subspace. Experimental results on the Yale and PIE face databases show that the proposed GHSA scheme achieves an improvement in face recognition accuracy when compared with conventional subspace approaches, and a further investigation shows that the proposed GHSA scheme has impressive robustness to noise.

Robust face recognition with partial occlusion, illumination variation and limited training data by optimal feature selection

This study investigates face recognition with partial occlusion, illumination variation and their combination, assuming no prior information about the mismatch, and limited training data for each person. The authors extend their previous posterior union model (PUM) to give a new method capable of dealing with all these problems. PUM is an approach for selecting the optimal local image features for recognition to improve robustness to partial occlusion. The extension is in two stages. First, authors extend PUM from a probability-based formulation to a similarity-based formulation, so that it operates with as little as one single training sample to offer robustness to partial occlusion. Second, they extend this new formulation to make it robust to illumination variation, and to combined illumination variation and partial occlusion, by a novel combination of multicondition relighting and optimal feature selection. To evaluate the new methods, a number of databases with various simulated and realistic occlusion/illumination mismatches have been used. The results have demonstrated the improved robustness of the new methods.

AdaBoost Multiple Feature Selection and Combination for Face Recognition

Gabor features have been recognized as one of the most successful face representations. Encouraged by the results given by this approach, other kind of facial representations based on Steerable Gaussian first order kernels and Harris corner detector are proposed in this paper. In order to reduce the high dimensional feature space, PCA and LDA techniques are employed. Once the features have been extracted, AdaBoost learning algorithm is used to select and combine the most representative features. The experimental results on XM2VTS database show an encouraging recognition rate, showing an important improvement with respect to face descriptors only based on Gabor filters.

A sequential algorithm for sparse support vector classifiers

Support vector machines (SVMs), though accurate, are not preferred in applications requiring high classification speed or when deployed in systems of limited computational resources, due to the large number of support vectors involved in the model. To overcome this problem we have devised a primal SVM method with the following properties: (1) it solves for the SVM representation without the need to invoke the representer theorem, (2) forward and backward selections are combined to approach the final globally optimal solution, and (3) a criterion is introduced for identification of support vectors leading to a much reduced support vector set. In addition to introducing this method the paper analyzes the complexity of the algorithm and presents test results on three public benchmark problems and a human activity recognition application. These applications demonstrate the effectiveness and efficiency of the proposed algorithm.


--------------------------------------------------------------------------------

Common-sense reasoning for human action recognition

This paper presents a novel method that leverages reasoning capabilities in a computer vision system dedicated to human action recognition. The proposed methodology is decomposed into two stages. First, a machine learning based algorithm - known as bag of words - gives a first estimate of action classification from video sequences, by performing an image feature analysis. Those results are afterward passed to a common-sense reasoning system, which analyses, selects and corrects the initial estimation yielded by the machine learning algorithm. This second stage resorts to the knowledge implicit in the rationality that motivates human behaviour. Experiments are performed in realistic conditions, where poor recognition rates by the machine learning techniques are significantly improved by the second stage in which common-sense knowledge and reasoning capabilities have been leveraged. This demonstrates the value of integrating common-sense capabilities into a computer vision pipeline. © 2012 Elsevier B.V. All rights reserved.

An Investigation Into Features For Multi-View Lipreading

For the first time in this paper we present results showing the effect of speaker head pose angle on automatic lip-reading performance over a wide range of closely spaced angles. We analyse the effect head pose has upon the features themselves and show that by selecting coefficients with minimum variance w.r.t. pose angle, recognition performance can be improved when train-test pose angles differ. Experiments are conducted using the initial phase of a unique multi view Audio-Visual database designed specifically for research and development of pose-invariant lip-reading systems. We firstly show that it is the higher order horizontal spatial frequency components that become most detrimental as the pose deviates. Secondly we assess the performance of different feature selection masks across a range of pose angles including a new mask based on Minimum Cross-Pose Variance coefficients. We report a relative improvement of 50% in Word Error Rate when using our selection mask over a common energy based selection during profile view lip-reading.

3D Morphable Model Construction for Robust Ear and Face Recognition

Recent work suggests that the human ear varies significantly between different subjects and can be used for identification. In principle, therefore, using ears in addition to the face within a recognition system could improve accuracy and robustness, particularly for non-frontal views. The paper describes work that investigates this hypothesis using an approach based on the construction of a 3D morphable model of the head and ear. One issue with creating a model that includes the ear is that existing training datasets contain noise and partial occlusion. Rather than exclude these regions manually, a classifier has been developed which automates this process. When combined with a robust registration algorithm the resulting system enables full head morphable models to be constructed efficiently using less constrained datasets. The algorithm has been evaluated using registration consistency, model coverage and minimalism metrics, which together demonstrate the accuracy of the approach. To make it easier to build on this work, the source code has been made available online.

An Event Driven Fusion Approach for Enjoyment Recognition in Real-time

Social signals and interpretation of carried information is of high importance in Human Computer Interaction. Often used for affect recognition, the cues within these signals are displayed in various modalities. Fusion of multi-modal signals is a natural and interesting way to improve automatic classification of emotions transported in social signals. Throughout most present studies, uni-modal affect recognition as well as multi-modal fusion, decisions are forced for fixed annotation segments across all modalities. In this paper, we investigate the less prevalent approach of event driven fusion, which indirectly accumulates asynchronous events in all modalities for final predictions. We present a fusion approach, handling short-timed events in a vector space, which is of special interest for real-time applications. We compare results of segmentation based uni-modal classification and fusion schemes to the event driven fusion approach. The evaluation is carried out via detection of enjoyment-episodes within the audiovisual Belfast Story-Telling Corpus.

New FMO Type to Flag ROI in H264/AVC

This paper presents a new type of Flexible Macroblock Ordering (FMO) type for the H.264 Advanced Video Coding (AVC) standard, which can more efficiently flag the position and shape of regions of interest (ROIs) in each frame. In H.264/AVC, 7 types of FMO have been defined, all of which are designed for error resilience. Most previous work related to ROI processing has adopted Type-2 (foreground & background), or Type-6 (explicit), to flag the position and shape of the ROI. However, only rectangular shapes are allowed in Type-2 and for non-rectangular shapes, the non-ROI macroblocks may be wrongly flagged as being within the ROI, which could seriously affect subsequent processing of the ROI. In Type-6, each macroblock in a frame uses fixed-length bits to indicate to its slice group. In general, each ROI is assigned to one slice group identity. Although this FMO type can more accurately flag the position and shape of the ROI, it incurs a significant bitrate overhead. The proposed new FMO type uses the smallest rectangle that covers the ROI to indicate its position and a spiral binary mask is employed within the rectangle to indicate the shape of the ROI. This technique can accurately flag the ROI and provide significantly savings in the bitrate overhead. Compared with Type-6, an 80% to 90% reduction in the bitrate overhead can be obtained while achieving the same accuracy.