S. Ali Etemad

An ant-inspired algorithm for detection of image edge features

This paper presents a technique inspired by swarm methodologies such as ant colony algorithms for processing simple and complicated images. It is shown that the proposed technique for image processing is capable of performing feature extraction for edge detection and segmentation, even in the presence of noise. Our proposed approach, Ant-based Correlation for Edge Detection (ACED), is tested on different samples and the results are compared to typical established non-swarm-based methods. The comparative analysis highlights the advantages of the proposed method which generates less distortion when noise is added to the test images. Both qualitative and quantitative evaluations support the claim, confirming the significance of our swarm-based method for image feature extraction and segmentation.

Robust Horizon Detection Using Segmentation for UAV Applications

A critical step in navigation of unmanned aerial vehicles is the detection of the horizon line. This information can be used for adjusting flight parameters as well as obstacle avoidance. In this paper, a fast and robust technique for precise detection of the horizon path is proposed. The method is based on existence of a unique light field that occurs in imagery where the horizon is viewed. This light field exists in different scenes including sea-sky, soil-sky, and forest-sky horizon lines. Our proposed approach employs segmentation of the scene and subsequent analysis of the image segments for extraction of the mentioned field and thus the horizon path. Through various experiments carried out on our own dataset and that of another previously published paper, we illustrate the significance and accuracy of this technique for various types of terrains from water to ground, and even snow-covered ground. Finally, it is shown that robust performance and accuracy, speed, and extraction of the path as curves (as opposed to a straight line which is resulted from many other approaches) are the benefits of our method.

Classification and translation of style and affect in human motion using RBF neural networks

Human motion can be carried out with a variety of different affects or styles such as happy, sad, energetic, and tired among many others. Modeling and classifying these styles, and more importantly, translating them from one sequence onto another has become a popular problem in the fields of graphics, multimedia, and human computer interaction. In this paper, radial basis functions (RBF) are used to model and extract stylistic and affective features from motion data. We demonstrate that using only a few basis functions per degree of freedom, successful modeling of styles in cycles of human walk can be achieved. Furthermore, we employ an ensemble of RBF neural networks to learn the affective/stylistic features following time warping and principal component analysis. The system learns the components and classifies stylistic motion sequences into distinct affective and stylistic classes. The system also utilizes the ensemble of neural networks to learn motion affects and styles such that it can translate them onto neutral input sequences. Experimental results along with both numerical and perceptual validations confirm the highly accurate and effective performance of the system.

Fast obstacle detection using targeted optical flow

This paper presents a new method for obstacle detection using optical flow. The method employs a highly efficient and accurate adaptive motion detection algorithm for determining the regions in the image which are more likely to contain obstacles. These regions then have optical flow performed on them. We call this method targeted optical flow. Targeted optical flow performs significantly faster compared to regular optical flow. We employ two types of optical flow to demonstrate the performance and speed increase of the proposed system. Finally, k-means clustering is employed for obstacle reconstruction. The system is designed for color videos for better performance. Several benchmark and recorded sequences have been used for testing the system.

Design and usability analysis of gesture-based control for common desktop tasks

We have designed and implemented a vision-based system capable of interacting with users natural arm and finger gestures. Using depth-based vision has reduced the effect of ambient disturbances such as noise and lighting condition. Various arm and finger gestures are designed and a system capable of detection and classification of gestures is developed and implemented. Finally the gesture recognition routine is linked to a simplified desktop for usability and human factor studies. Several factors such as precision, efficiency, ease-of-use, pleasure, fatigue, naturalness, and overall satisfaction are investigated in detail. Through different simple and complex tasks, it is concluded that finger-based inputs are superior to arm-based ones in the long run. Furthermore, it is shown that arm gestures cause more fatigue and appear less natural than finger gestures. However, factors such as time, overall satisfaction, and easiness were not affected by selecting one over the other.

Gamification of Exercise and Fitness using Wearable Activity Trackers

Wearable technologies are a growing industry with significant potential in different aspects of health and fitness. Gamification of health and fitness, on the other hand, has recently become a popular field of research. Accordingly, we believe that wearable devices have the potential to be utilized towards gamification of fitness and exercise. In this paper, we first review several popular activity tracking wearable devices, their characteristics and specifications, and their application programming interface (API) capabilities and availabilities, which will enable them to be employed by third party developers for the purpose at hand. The feasibility and potential advantages of utilizing wearables for gamification of health and fitness are then discussed. Finally, we develop a pilot prototype as a case-study for this concept, and perform preliminary user studies which will help further explore the proposed concept.

Usability and Motivational Effects of a Gamified Exercise and Fitness System Based on Wearable Devices

Gamification of exercise has become a popular topic due to its motivational and engagement effects, which intends to result in increased exercise, activity, health, and fitness for everyday users. At the same time, wearable technologies have become a fast growing industry, providing consumers the ability to conveniently track their state of health and fitness efforts. Consequently, we propose that off-the-shelf commercial wearable technologies have great potential to be applied in gamification of fitness and exercise. In this paper, we utilize this concept through the design and implementation of a smartphone game application which uses wearable devices as input systems. The game supports two different wearable devices which are commercially available to end-users, and three types of activities were included in the gamified experience. These activities are performed as inputs to our mobile game. User tests evaluate the effectiveness of the combined use of games and wearable devices in promoting exercise. The usability of our proposed approach and effects of different factors within the system are also evaluated.

Correlation-optimized time warping for motion

Retrieval and comparative editing/modeling of motion data require temporal alignment. In other words, for such processes to perform accurately, critical features of motion sequences need to occur simultaneously. In this paper, we propose correlation-optimized time warping (CoTW) for aligning motion data. CoTW utilizes a correlation-based objective function for characterizing alignment. The method solves an optimization problem to determine the optimum warping degree for different segments of the sequence. Using segment-wise interpolated warping, smooth motion trajectories are achieved that can be readily used for animation. Our method allows for manual tuning of the parameters, resulting in high customizability with respect to the number of actions in a single sequence as well as spatial regions of interest within the character model. Moreover, measures are taken to reduce distortion caused by over-warping. The framework also allows for automatic selection of an optimum reference when multiple sequences are available. Experimental results demonstrate the very accurate performance of CoTW compared to other techniques such as dynamic time warping, derivative dynamic time warping and canonical time warping. The mentioned customization capabilities are also illustrated.

Additivity in perception of affect from limb motion.

In this study, the notion of additivity in perception of affect from limb motion is investigated. Specifically, we examine whether the impact of multiple limbs in perception of affect is equal to the sum of the impacts of each individual limb. Several neutral, happy, and sad walking sequences are first aligned and averaged. Four distinct body regions or limbs are defined for this study: arms and hands, legs and feet, head and neck, and torso. The three average walks are used to create the stimuli. The motion of each limb and combination of limbs from the neutral sequence are replaced with those of the happy and sad sequences. Through collecting perceptual ratings for when individual limbs contain affective features, and comparing the sums of these ratings to instances where multiple limbs of the body simultaneously contain affective features, additivity is investigated. We find that while the results are highly correlated, additivity does not hold in the classical sense. Based on the results, a mathematical model is proposed for describing the observed relationship.

A Customizable Time Warping Method for Motion Alignment

This paper presents Correlation-optimized Time Warping (CoTW) for aligning motion sequences. The proposed method maximizes an objective function based on the correlation of the two sequences. There are several parameters involved in the process, which can be computationally optimized or manually customized. Customization can take place based on the number and/or nature of actions in the sequences. CoTW shows robust performance for aligning simple gait sequences as well as sequences containing several different actions.