Javed Ahmed

Action MACH a spatio-temporal Maximum Average Correlation Height filter for action recognition

In this paper we introduce a template-based method for recognizing human actions called action MACH. Our approach is based on a maximum average correlation height (MACH) filter. A common limitation of template-based methods is their inability to generate a single template using a collection of examples. MACH is capable of capturing intra-class variability by synthesizing a single Action MACH filter for a given action class. We generalize the traditional MACH filter to video (3D spatiotemporal volume), and vector valued data. By analyzing the response of the filter in the frequency domain, we avoid the high computational cost commonly incurred in template-based approaches. Vector valued data is analyzed using the Clifford Fourier transform, a generalization of the Fourier transform intended for both scalar and vector-valued data. Finally, we perform an extensive set of experiments and compare our method with some of the most recent approaches in the field by using publicly available datasets, and two new annotated human action datasets which include actions performed in classic feature films and sports broadcast television.

Correlation, Kalman filter and adaptive fast mean shift based heuristic approach for robust visual tracking

Correlation tracker is computation intensive (if the search space or the template is large), has template drift problem, and may fail in case of fast maneuvering target, rapid changes in its appearance, occlusion suffered by it and clutter in the scene. Kalman filter can predict the target coordinates in the next frame, if the measurement vector is supplied to it by a correlation tracker. Thus, a relatively small search space can be determined where the probability of finding the target in the next frame is high. This way, the tracker can become fast and reject the clutter, which is outside the search space in the scene. However, if the tracker provides wrong measurement vector due to the clutter or the occlusion inside the search region, the efficacy of the filter is significantly deteriorated. Mean-shift tracker is fast and has shown good tracking results in the literature, but it may fail when the histograms of the target and the candidate region in the scene are similar (even when their appearance is different). In order to make the overall visual tracking framework robust to the mentioned problems, we propose to combine the three approaches heuristically, so that they may support each other for better tracking results. Furthermore, we present novel method for (1) appearance model updating which adapts the template according to rate of appearance change of target, (2) adaptive threshold for similarity measure which uses the variable threshold for each forthcoming image frame based on current frame peak similarity value, and (3) adaptive kernel size for fast mean-shift algorithm based on varying size of the target. Comparison with nine state-of-the-art tracking algorithms on eleven publically available standard dataset shows that the proposed algorithm outperforms the other algorithms in most of the cases.

Target Tracking in an Image Sequence Using Wavelet Features and a Neural Network

In this paper, we present a comprehensive study to design an artificial neural network (ANN) for tracking a target in an image sequence. The proposed ANN architecture is a single-hidden-layer back-propagation neural network (BPNN), in which the sigmoid and the linear activation functions are used for its hidden and output layers, respectively. The features used for the input layer of the BPNN are 4th level Daubechiess wavelet decomposition coefficients corresponding to the input image. Performances of dbl, db2, db3, and db4 wavelet features are compared. The object, which is tracked for the purpose of demonstration, is a specific airplane. However, the proposed ANN model can be trained to track any other object of interest. The trained ANN has been simulated and tested on the training and testing datasets. The tracking error is analyzed with post-regression analysis tool, which finds the correlation among the calculated coordinates and the correct coordinates of the object in the image. The promising results of the presented computer simulation and analysis show that the proposed target tracking technique exploiting the powers of ANN and wavelet transform is quite plausible and significantly robust.

Robust Edge-Enhanced Fragment Based Normalized Correlation Tracking in Cluttered and Occluded Imagery

Correlation trackers are in use for the past four decades. Edge based correlation tracking algorithms have proved their strength for long term tracking, but these algorithms suffer from two major problems: clutter and slow occlusion. Thus, there is a requirement to improve the confidence measure regarding target and non-target object. In order to solve these problems, we present an “Edge Enhanced Fragment Based Normalized Correlation (EEFNC)” algorithm, in which we: (1) divide the target template into nine non-overlapping fragments after edge-enhancement, (2) correlate each fragment with the corresponding fragment of the template-size section in the search region, and (3) achieve the final similarity measure by averaging the correlation values obtained for every fragment. A fragment level template updating method is also proposed to make the template adaptive to the variation in the shape and appearance of the object in motion. We provide the experimental results which show that the proposed technique outperforms the recent Edge-Enhanced Normalized Correlation (EENC) tracking algorithm in occlusion and clutter.

Visual object tracking--classical and contemporary approaches

Visual object tracking (VOT) is an important subfield of computer vision. It has widespread application domains, and has been considered as an important part of surveillance and security system. VOA facilitates finding the position of target in image coordinates of video frames.While doing this, VOA also faces many challenges such as noise, clutter, occlusion, rapid change in object appearances, highly maneuvered (complex) object motion, illumination changes. In recent years, VOT has made significant progress due to availability of low-cost high-quality video cameras as well as fast computational resources, and many modern techniques have been proposed to handle the challenges faced by VOT. This article introduces the readers to 1) VOT and its applications in other domains, 2) different issues which arise in it, 3) various classical as well as contemporary approaches for object tracking, 4) evaluation methodologies for VOT, and 5) online resources, i.e., annotated datasets and source code available for various tracking techniques.

Improved Phase Correlation Matching

Phase correlation based template matching is an efficient tool for translation estimation which is in turn required for the image registration and the object tracking applications. When a template of an object is phase correlated with the search image, the resulting correlation surface is supposed to contain a sharp peak corresponding to the location of the object in the search image. However, the resulting surface also contains various falsepeaks which are sometimes higherin magnitude than the truepeak. In order to solve the problem, we present an efficient and effective preprocessing technique that extends the images with new pixels having decaying values. The technique is compared with two recent methods on cluttered, noisy, blurred, and slightly rotated scenes. The results show that the proposed method outperforms both of them, especially when the object is away from the central region in the image.

Wigner-Ville distribution of a type of non-stationary mobile Rayleigh fading channels

The classical Clarke model of mobile radio reception assumes a constant mobile velocity. We, in this paper, relax the assumption of constant mobile velocity to allow the mobile to have constant acceleration and derive expression for the non-stationary autocorrelation function of the channel process in general 2-dimensional (2D) scattering environments. Under suitable assumptions, an expression for Wigner-Ville spectrum is obtained in isotropic scattering environment which suggests that the Wigner-Ville spectrum is a natural generalization of the Clarkes model to constant mobile acceleration scenario.

Fusion of color and infrared images using gradient transfer and total variation minimization

The objective of this paper is to propose an algorithm for fusion of images of two different image modalities, i.e., color visual image and its corresponding infrared (IR) image. Fusion method is based on l1 total variation minimization technique, and it combines appearance detail and thermal information for a scene using visual and IR images, respectively. Moreover, the proposed method maintains the natural color of the visual image in the fused image. Normally, IR image highlights the camouflage or hidden target in a scene due to its thermal variations from other objects. Therefore, the fused image is more detailed single image than its constituent images, and it reveals the concealed information which is not visible in input images. Experimental results show the effectiveness of the algorithm.

Computer vision based intruder detection framework (CV-IDF)

Surveillance has become a major issue in recent years after food and clothing, the focus is how to prevent sudden attacks to secure our lives. In this paper, we propose a robust algorithm which aims at detecting and tracking multiple intruders in a forbidden area. Significant additions of our paper are that we propose a reliable and efficient framework which uses background subtraction method using Gaussian mixture models followed by Kalman filter to detect intruders in a restricted area, which reduces false alarms. It tracks multiple intruders moving inside a restricted region using Kalman filter, data associations of predictions obtained by Kalman filter and track of individual intruder is achieved by Hungarian algorithm. Alarms are generated after getting detections to notify security staff for taking appropriate actions.

A new template updating method for correlation tracking

One of the most challenging problems faced by tracking algorithms is the issue of template drift. For robust object tracking, template should be adaptive enough to incorporate maximum changes of target appearance, and at same time it should be restrictive enough to reject any background information entering into its model so that drifting of template may be avoided. The existing template updating methods does not take into account the actual appearance changes of the targets, therefore, they are not much effective against template drift. This paper proposes a new template updating method for correlation based tracking algorithms which updates the template according to rate of change in appearance of the target. Moreover, the method is capable to revert the template near to some previous better representation if recent updating is incorrect. Thus, the proposed algorithm helps to overcome the problems of template drift especially during slow occurring occlusion and complex (e.g., out-of-plane) motion of the target. Experimental results and comparison with other algorithms on different publicly available challenging videos prove the efficacy of the algorithm.