Using Biologically Inspired Features for Face Processing

In this paper, we show that a new set of visual features, derived from a feed-forward model of the primate visual object recognition pathway proposed by Riesenhuber and Poggio (R&P Model) (Nature Neurosci. 2(11):1019–1025, 1999) is capable of matching the performance of some of the best current representations for face identification and facial expression recognition. Previous work has shown that the Riesenhuber and Poggio Model features can achieve a high level of performance on object recognition tasks (Serre, T., et al. in IEEE Comput. Vis. Pattern Recognit. 2:994–1000, 2005). Here we modify the R&P model in order to create a new set of features useful for face identification and expression recognition. Results from tests on the FERET, ORL and AR datasets show that these features are capable of matching and sometimes outperforming other top visual features such as local binary patterns (Ahonen, T., et al. in 8th European Conference on Computer Vision, pp. 469–481, 2004) and histogram of gradient features (Dalal, N., Triggs, B. in International Conference on Computer Vision & Pattern Recognition, pp. 886–893, 2005). Having a model based on shared lower level features, and face and object recognition specific higher level features, is consistent with findings from electrophysiology and functional magnetic resonance imaging experiments. Thus, our model begins to address the complete recognition problem in a biologically plausible way.     

Feature subset selection based on fuzzy entropy measures for handling classification problems

In this paper, we present a new method for dealing with feature subset selection based on fuzzy entropy measures for handling classification problems. First, we discretize numeric features to construct the membership function of each fuzzy set of a feature. Then, we select the feature subset based on the proposed fuzzy entropy measure focusing on boundary samples. The proposed method can select relevant features to get higher average classification accuracy rates than the ones selected by the MIFS method (Battiti, R. in IEEE Trans. Neural Netw. 5(4):537–550, 1994), the FQI method (De, R.K., et al. in Neural Netw. 12(10):1429–1455, 1999), the OFEI method, Dong-and-Kothari’s method (Dong, M., Kothari, R. in Pattern Recognit. Lett. 24(9):1215–1225, 2003) and the OFFSS method (Tsang, E.C.C., et al. in IEEE Trans. Fuzzy Syst. 11(2):202–213, 2003).

Linear Time Algorithms for Exact Distance Transform

In 2003, Maurer et al. (IEEE Trans. Pattern Anal. Mach. Intell. 25:265–270, 2003) published a paper describing an algorithm that computes the exact distance transform in linear time (with respect to image size) for the rectangular binary images in the k-dimensional space ℝ ^k and distance measured with respect to L _p -metric for 1≤p≤∞, which includes Euclidean distance L ₂. In this paper we discuss this algorithm from theoretical and practical points of view. On the practical side, we concentrate on its Euclidean distance version, discuss the possible ways of implementing it as signed distance transform, and experimentally compare implemented algorithms. We also describe the parallelization of these algorithms and discuss the computational time savings associated with them. All these implementations will be made available as a part of the CAVASS software system developed and maintained in our group (Grevera et al. in J. Digit. Imaging 20:101–118, 2007). On the theoretical side, we prove that our version of the signed distance transform algorithm, GBDT, returns the exact value of the distance from the geometrically defined object boundary. We provide a complete proof (which was not given of Maurer et al. (IEEE Trans. Pattern Anal. Mach. Intell. 25:265–270, 2003) that all these algorithms work correctly for L _p -metric with 1<p<∞. We also point out that the precise form of the algorithm from Maurer et al. (IEEE Trans. Pattern Anal. Mach. Intell. 25:265–270, 2003) is not well defined for L ₁ and L _∞ metrics. In addition, we show that the algorithm can be used to find, in linear time, the exact value of the diameter of an object, that is, the largest possible distance between any two of its elements.     

Population Shape Regression from Random Design Data

Regression analysis is a powerful tool for the study of changes in a dependent variable as a function of an independent regressor variable, and in particular it is applicable to the study of anatomical growth and shape change. When the underlying process can be modeled by parameters in a Euclidean space, classical regression techniques (Hardle, Applied Nonparametric Regression, 1990; Wand and Jones, Kernel Smoothing, 1995) are applicable and have been studied extensively. However, recent work suggests that attempts to describe anatomical shapes using flat Euclidean spaces undermines our ability to represent natural biological variability (Fletcher et al., IEEE Trans. Med. Imaging 23(8), 995–1005, 2004; Grenander and Miller, Q. Appl. Math. 56(4), 617–694, 1998).     

A comparative study of existing metrics for 3D-mesh segmentation evaluation

In this paper, we present an extensive experimental comparison of existing similarity metrics addressing the quality assessment problem of mesh segmentation. We introduce a new metric, named the 3D Normalized Probabilistic Rand Index (3D-NPRI), which outperforms the others in terms of properties and discriminative power. This comparative study includes a subjective experiment with human observers and is based on a corpus of manually segmented models. This corpus is an improved version of our previous one (Benhabiles et al. in IEEE International Conference on Shape Modeling and Application (SMI), 2009). It is composed of a set of 3D-mesh models grouped in different classes associated with several manual ground-truth segmentations. Finally the 3D-NPRI is applied to evaluate six recent segmentation algorithms using our corpus and the Chen et al.’s (ACM Trans. Graph. (SIGGRAPH), 28(3), 2009) corpus.     

Younger and older users׳ recognition of virtual agent facial expressions

As technology advances, robots and virtual agents will be introduced into the home and healthcare settings to assist individuals, both young and old, with everyday living tasks. Understanding how users recognize an agent׳s social cues is therefore imperative, especially in social interactions. Facial expression, in particular, is one of the most common non-verbal cues used to display and communicate emotion in on-screen agents (Cassell et al., 2000). Age is important to consider because age-related differences in emotion recognition of human facial expression have been supported (Ruffman et al., 2008), with older adults showing a deficit for recognition of negative facial expressions. Previous work has shown that younger adults can effectively recognize facial emotions displayed by agents (Bartneck and Reichenbach, 2005, Courgeon et al., 2009, Courgeon et al., 2011, Breazeal, 2003); however, little research has compared in-depth younger and older adults’ ability to label a virtual agent׳s facial emotions, an import consideration because social agents will be required to interact with users of varying ages. If such age-related differences exist for recognition of virtual agent facial expressions, we aim to understand if those age-related differences are influenced by the intensity of the emotion, dynamic formation of emotion (i.e., a neutral expression developing into an expression of emotion through motion), or the type of virtual character differing by human-likeness. Study 1 investigated the relationship between age-related differences, the implication of dynamic formation of emotion, and the role of emotion intensity in emotion recognition of the facial expressions of a virtual agent (iCat). Study 2 examined age-related differences in recognition expressed by three types of virtual characters differing by human-likeness (non-humanoid iCat, synthetic human, and human). Study 2 also investigated the role of configural and featural processing as a possible explanation for age-related differences in emotion recognition. First, our findings show age-related differences in the recognition of emotions expressed by a virtual agent, with older adults showing lower recognition for the emotions of anger, disgust, fear, happiness, sadness, and neutral. These age-related difference might be explained by older adults having difficulty discriminating similarity in configural arrangement of facial features for certain emotions; for example, older adults often mislabeled the similar emotions of fear as surprise. Second, our results did not provide evidence for the dynamic formation improving emotion recognition; but, in general, the intensity of the emotion improved recognition. Lastly, we learned that emotion recognition, for older and younger adults, differed by character type, from best to worst: human, synthetic human, and then iCat. Our findings provide guidance for design, as well as the development of a framework of age-related differences in emotion recognition.     

Comparison of parallel multi-objective approaches to protein structure prediction

Protein structure prediction (PSP) is an open problem with many useful applications in disciplines such as medicine, biology and biochemistry. As this problem presents a vast search space and the analysis of each protein structure requires a significant amount of computing time, it is necessary to take advantage of high-performance parallel computing platforms as well as to define efficient search procedures in the space of possible protein conformations. In this paper we compare two parallel procedures for PSP which are based on different multi-objective optimization approaches, i.e. PAES (Knowles and Corne in Proc. Congr. Evol. Comput. 1:98–105, 1999) and NSGA2 (Deb et al. in IEEE Trans. Evol. Comput. 6:182–197, 2002). Although both procedures include techniques to take advantage of known protein structures and strategies to simplify the search space through the so-called rotamer library and adaptive mutation operators, they present different profiles with respect to their implicit parallelism.     

An earlier time for inserting and/or accelerating tasks

In a periodic real-time system scheduled by the EDF (Earliest Deadline First) algorithm (Liu and Layland, J. ACM 20(1), 40–61, 1973; Barauh, Proc. of the 27th IEEE International Real-Time Systems Symposium, 379–387, 2006; Buttazzo, J. Real-Time Syst. 29(1), 5–26, 2005), when new tasks have to be inserted into the system at run-time and/or current tasks request to increase their rates in response to internal or external events, the new sum of the utilizations after the insertion and/or acceleration should be limited, otherwise, one or more current tasks should usually be compressed (their periods being prolonged) in order to avoid overload. Buttazzo offered a time from which on this kind of adjustment can be done without causing any deadline miss in the system (Buttazzo et al., IEEE Trans. Comput. 51(3), 289–302, 2002). It is, however, not early enough. In this paper, an earlier time is given and formally proved.

An improved data hiding approach for polygon meshes

We present an improved technique for data hiding in polygonal meshes, which is based on the work of Bogomjakov et al. (Comput. Graph. Forum 27(2):637–642, 2008). Like their method, we use an arrangement on primitives relative to a reference ordering to embed a message. But instead of directly interpreting the index of a primitive in the reference ordering as the encoded/decoded bits, our method slightly modifies the mapping so that our modification doubles the chance of encoding an additional bit compared to Bogomjakov et al.’s (Comput. Graph. Forum 27(2):637–642, 2008). We illustrate the inefficiency in the original mapping of Bogomjakov et al. (Comput. Graph. Forum 27(2):637–642, 2008) with an intuitive representation using a binary tree.     

Ontology-driven coordination model for multiagent-based mobile workforce brokering systems

Coordination has been recognized by many researchers as the most important feature of multi-agent systems. Coordination is defined as managing interdependencies amongst activities (Malone and Crowston in ACM Comput. Surv. 26(1):87–119, 1994). The traditional approach of implementing a coordination mechanism is to hard-wire it into a coordination system at design time. However, in dynamic and open environments, many attributes of the system cannot be accurately identified at the design time. Therefore, dynamic coordination, capable of coordinating activities at run-time, has emerged. On the other hand, a successful dynamic coordination model for multi-agent systems requires knowledge sharing as well as common vocabulary. Therefore, an ontological approach is an appropriate way in proposing dynamic coordination models for multi-agent systems. In this paper, an Ontology-Driven Dynamic Coordination Model (O-DC) for Multiagent-Based Mobile Workforce Brokering Systems (MWBS) (Mousavi et al. in Int. J. Comput. Sci. 6:(5):557–565, 2010; Mousavi et al. in Proceedings of 4th IEEE international symposium on information technology, ITSim’10, Kuala Lumpur, Malaysia, 15–17 June 2010, vol. 3, pp. 1416–1421, 2010; Mousavi and Nordin in Proceedings of the IEEE international conference on electrical engineering and informatics, Bandung, Indonesia, 17–19 June 2007, pp. 294–297, 2007) is proposed and formulated. Subsequently, the applicability of O-DC is examined via simulation based on a real-world scenario.     

Facial expression recognition using kernel canonical correlation analysis (KCCA)

In this correspondence, we address the facial expression recognition problem using kernel canonical correlation analysis (KCCA). Following the method proposed by Lyons et al. and Zhang et al. , we manually locate 34 landmark points from each facial image and then convert these geometric points into a labeled graph (LG) vector using the Gabor wavelet transformation method to represent the facial features. On the other hand, for each training facial image, the semantic ratings describing the basic expressions are combined into a six-dimensional semantic expression vector. Learning the correlation between the LG vector and the semantic expression vector is performed by KCCA. According to this correlation, we estimate the associated semantic expression vector of a given test image and then perform the expression classification according to this estimated semantic expression vector. Moreover, we also propose an improved KCCA algorithm to tackle the singularity problem of the Gram matrix. The experimental results on the Japanese female facial expression database and the Ekman's "Pictures of Facial Affect" database illustrate the effectiveness of the proposed method.     

Constructing a pedestrian recognition system with a public open database, without the necessity of re-training: an experimental study

Pedestrian recognition is one of the basic elements of an active pedestrian protection system. Especially, there have been various researches to develop vision-based pedestrian classifiers. Recently, Munder and Gavrila (IEEE Trans Pattern Anal Mach Intell 28(11):1863–1868, 2006) opened their DaimlerChrysler (DCX) pedestrian image database and tried to provide an objective comparison between popular features and classifiers. After their publication, objective performance comparison of features, classifiers and architectures became possible. This paper reports four experimental results with the DCX database. First, the Gabor filter bank-based feature is competent in pedestrian recognition. Second, optimization of a classifier using performance estimator greatly enhances the performance of the resultant classifier. Third, once the imaging system uses histogram equalization and the same capturing method as a database, a practical pedestrian recognition system can be constructed with a public open database, without the necessity of re-training using the new database acquired with the actual imaging system. Fourth, a posteriori probability-based post-processing increases the recognition rate of consecutive image sequencing while maintaining a false positive rate.     

Adaptive listen for energy-efficient medium access control in wireless sensor networks

Wireless Sensor Networks (WSN) have nodes that are small in size and are powered by small batteries having very limited amount of energy. In most applications of WSN, the nodes in the network remain inactive for long periods of time, and intermittently they become active on sensing any change in the environment. The data sensed by the different nodes are sent to the sink node. In contrast to other infrastructure-based wireless networks, higher throughput, lower latency and per-node fairness in WSN are imperative, but their importance is subdued when compared to energy consumption. In this work, we have regarded the amount of energy consumption in the nodes to be of primary concern, while throughput and latency in the network to be secondary. We have proposed a protocol for energy-efficient adaptive listen for medium access control in WSN. Our protocol adaptively changes the slot-time, which is the time of each slot in the contention window. This correspondingly changes the cycle-time, which is the sum of the listen-time and the sleep-time of the sensors, while keeping the duty-cycle, which is the ratio between the listen-time and the cycle-time, constant. Using simulation experiments, we evaluated the performance of the proposed protocol, compared with the popular Sensor Medium Access Control (SMAC) (Ye et al. IEEE/ACM Trans Netw 12(3):493–506, 39) protocol. The results we obtained show a prominent decrease in the energy consumption at the nodes in the proposed protocol over the existing SMAC protocol, at the cost of decreasing the throughput and increasing the latency in the network. Although such an observation is not perfectly what is ideally desired, given the very limited amount of energy with which the nodes in a WSN operate, we advocate that increasing the energy efficiency of the nodes, thereby increasing the network lifetime in WSN, is a more important concern compared to throughput and latency. Additionally, similar observations relating energy efficiency, network lifetime, throughput and latency exist in many other existing protocols, including the popular SMAC protocol (Ye et al. IEEE/ACM Trans Netw 12(3):493–506, 39).     

The effects of scent and game play experience on memory of a virtual environment

Scent has been well documented as having significant effects on emotion (Alaoui-Ismaili in Physiol Behav 62(4):713–720, 1997; Herz et al. in Motiv Emot 28(4):363–383, 2004), learning (Smith et al. in Percept Mot Skills 74(2):339–343, 1992; Morgan in Percept Mot Skills 83(3)(2):1227–1234, 1996), memory (Herz in Am J Psychol 110(4):489–505, 1997) and task performance (Barker et al. in Percept Mot Skills 97(3)(1):1007–1010, 2003). This paper describes an experiment in which environmentally appropriate scent was presented as an additional sensory modality consistent with other aspects of a virtual environment called DarkCon. Subjects’ game play habits were recorded as an additional factor for analysis. Subjects were randomly assigned to receive scent during the VE, and/or afterward during a task of recall of the environment. It was hypothesized that scent presentation during the VE would significantly improve recall, and that subjects who were presented with scent during the recall task, in addition to experiencing the scented VE, would perform the best on the recall task. Skin-conductance was a significant predictor of recall, over and above experimental groups. Finally, it was hypothesized that subjects’ game play habits would affect both their behavior in and recall of the environment. Results are encouraging to the use of scent in virtual environments, and directions for future research are discussed. The project described herein has been sponsored by the US Army Research, Development, and Engineering Command (RDECOM). Statements and opinions expressed do not necessarily reflect the position or the policy of the US Government; no official endorsement should be inferred.     

Facial expression recognition based on local region specific features and support vector machines

Facial expressions are one of the most powerful, natural and immediate means for human being to communicate their emotions and intensions. Recognition of facial expression has many applications including human-computer interaction, cognitive science, human emotion analysis, personality development etc. In this paper, we propose a new method for the recognition of facial expressions from single image frame that uses combination of appearance and geometric features with support vector machines classification. In general, appearance features for the recognition of facial expressions are computed by dividing face region into regular grid (holistic representation). But, in this paper we extracted region specific appearance features by dividing the whole face region into domain specific local regions. Geometric features are also extracted from corresponding domain specific regions. In addition, important local regions are determined by using incremental search approach which results in the reduction of feature dimension and improvement in recognition accuracy. The results of facial expressions recognition using features from domain specific regions are also compared with the results obtained using holistic representation. The performance of the proposed facial expression recognition system has been validated on publicly available extended Cohn-Kanade (CK+) facial expression data sets.     

A Tensor Approximation Approach to Dimensionality Reduction

Dimensionality reduction has recently been extensively studied for computer vision applications. We present a novel multilinear algebra based approach to reduced dimensionality representation of multidimensional data, such as image ensembles, video sequences and volume data. Before reducing the dimensionality we do not convert it into a vector as is done by traditional dimensionality reduction techniques like PCA. Our approach works directly on the multidimensional form of the data (matrix in 2D and tensor in higher dimensions) to yield what we call a Datum-as-Is representation. This helps exploit spatio-temporal redundancies with less information loss than image-as-vector methods. An efficient rank-R tensor approximation algorithm is presented to approximate higher-order tensors. We show that rank-R tensor approximation using Datum-as-Is representation generalizes many existing approaches that use image-as-matrix representation, such as generalized low rank approximation of matrices (GLRAM) (Ye, Y. in Mach. Learn. 61:167–191, 2005), rank-one decomposition of matrices (RODM) (Shashua, A., Levin, A. in CVPR’01: Proceedings of the 2001 IEEE computer society conference on computer vision and pattern recognition, p. 42, 2001) and rank-one decomposition of tensors (RODT) (Wang, H., Ahuja, N. in ICPR ’04: ICPR ’04: Proceedings of the 17th international conference on pattern recognition (ICPR’04), vol. 1, pp. 44–47, 2004). Our approach yields the most compact data representation among all known image-as-matrix methods. In addition, we propose another rank-R tensor approximation algorithm based on slice projection of third-order tensors, which needs fewer iterations for convergence for the important special case of 2D image ensembles, e.g., video. We evaluated the performance of our approach vs. other approaches on a number of datasets with the following two main results. First, for a fixed compression ratio, the proposed algorithm yields the best representation of image ensembles visually as well as in the least squares sense. Second, proposed representation gives the best performance for object classification. A shorter version of this paper was published at IEEE CVPR 2005 (Wang and Ahuja 2005).     

A Linear-Time Algorithm for Symmetric Convex Drawings of Internally Triconnected Plane Graphs

We study a crossing minimization problem of drawing a bipartite graph with a radial drawing of two orbits. Radial drawings are one of well-known drawing conventions in social network analysis and visualization, in particular, displaying centrality indices of actors (Wasserman and Faust, Social Network Analysis: Methods and Applications. Cambridge University Press, Cambridge, 1994). The main problem in this paper is called the one-sided radial crossing minimization, if the positions of vertices in the outer orbit are fixed. The problem is known to be NP-hard (Bachmaier, IEEE Trans. Vis. Comput. Graph. 13, 583–594, 2007), and a number of heuristics are available (Bachmaier, IEEE Trans. Vis. Comput. Graph. 13, 583–594, 2007). However, there is no approximation algorithm for the crossing minimization problem in radial drawings. We present the first polynomial time constant-factor approximation algorithm for the one-sided radial crossing minimization problem.     

Ordered Hypothesis Machines

Stack Filters are a class of non-linear filter typically used for noise suppression. Advantages of Stack Filters are their generality and the existence of efficient optimization algorithms under mean absolute error (Wendt et al. in IEEE Trans. Acoust. Speech Signal Process. 34:898–910, 1986). In this paper we describe our recent efforts to use the class of Stack Filters for classification problems. This leads to a novel class of continuous domain classifiers which we call Ordered Hypothesis Machines (OHM). We develop convex optimization based learning algorithms for Ordered Hypothesis Machines and highlight their relationship to Support Vector Machines and Nearest Neighbor classifiers. We report on the performance on synthetic and real-world datasets including an application to change detection in remote sensing imagery. We conclude that OHM provides a novel way to reduce the number of exemplars used in Nearest Neighbor classifiers and achieves competitive performance to the more computationally expensive K-Nearest Neighbor method.     

Improved Approximation Results for the Minimum Energy Broadcasting Problem

In this paper we present new results on the performance of the Minimum Spanning Tree heuristic for the Minimum Energy Broadcast Routing (MEBR) problem. We first prove that, for any number of dimensions d≥2, the approximation ratio of the heuristic does not increase when the power attenuation coefficient α, that is the exponent to which the coverage distance must be raised to give the emission power, grows. Moreover, we show that, for any fixed instance, as a limit for α going to infinity, the ratio tends to the lower bound of Clementi et al. (Proceedings of the 18th annual symposium on theoretical aspects of computer science (STACS), pp. 121–131, 2001), Wan et al. (Wirel. Netw. 8(6):607–617, 2002) given by the d-dimensional kissing number, thus closing the existing gap between the upper and the lower bound. We then introduce a new analysis allowing to establish a 7.45-approximation ratio for the 2-dimensional case, thus significantly decreasing the previously known 12 upper bound (Wan et al. in Wirel. Netw. 8(6):607–617, 2002) (actually corrected to 12.15 in Klasing et al. (Proceedings of the 3rd IFIP-TC6 international networking conference, pp. 866–877, 2004)). Finally, we extend our analysis to any number of dimensions d≥2 and any α≥d, obtaining a general approximation ratio of 3 ^d −1, again independent of α. The improvements of the approximation ratios are specifically significant in comparison with the lower bounds given by the kissing numbers, as these grow at least exponentially with respect to d. The research was partially funded by the European project COST Action 293, “Graphs and Algorithms in Communication Networks” (GRAAL). Preliminary version of this paper appeared in Flammini et al. (Proceedings of ACM joint workshop on foundations of mobile computing (DIALM-POMC), pp. 85–91, 2004).     

Distributed SAT

We present DPLL ABT, a distributed Satisfiability solver (SAT) (Ansótegui and Manyà in IberoAm J Artif Intell 7(20):43–56, 2003) designed to solve distributed SAT problem instances. Since SAT is a particular case of constraint satisfaction, we propose a solving method based on the Asynchronous Backtracking algorithm (ABT) (Yokoo et al. in IEEE Trans Knowl Data Eng 10(5):673–685, 1998) developed for distributed constraint reasoning. In addition, we have applied the Davis-Putnam procedure (DPLL) in every agent, plus the minimum conflict heuristic in case DPLL does not detect any inconsistency. The resulting algorithm improves the performance in terms of communication cost and computational effort versus the basic ABT. The SAT instance is distributed into agents, which cooperate to solve SAT instances just sharing the minimum information. We also present the experimental results that demonstrate the performance of the method in terms of communication and execution time comparing the performance with the basic ABT algorithm.