Du Yong Kim

Visual tracking via adaptive tracker selection with multiple features

In this paper, a robust visual tracking method is proposed to track an object in dynamic conditions that include motion blur, illumination changes, pose variations, and occlusions. To cope with these challenges, multiple trackers with different feature descriptors are utilized, and each of which shows different level of robustness to certain changes in an objects appearance. To fuse these independent trackers, we propose two configurations, tracker selection and interaction. The tracker interaction is achieved based on a transition probability matrix (TPM) in a probabilistic manner. The tracker selection extracts one tracking result from among multiple tracker outputs by choosing the tracker that has the highest tracker probability. According to various changes in an objects appearance, the TPM and tracker probability are updated in a recursive Bayesian form by evaluating each trackers reliability, which is measured by a robust tracker likelihood function (TLF). When the tracking in each frame is completed, the estimated objects state is obtained and fed into the reference update via the proposed learning strategy, which retains the robustness and adaptability of the TLF and multiple trackers. The experimental results demonstrate that our proposed method is robust in various benchmark scenarios.

Stem cell migration and mechanotransduction on linear stiffness gradient hydrogels

Significance Mechanobiology is receiving an increasing amount of focus, but the mechanics of cell-substrate behavior are often neglected in cell biology. As such, novel materials and systems that are simple to build and share in a nonengineering laboratory are sorely needed. We have fabricated gradient hydrogels with continuous linear gradients above and below the durotactic threshold, making it possible to pinpoint optimal stiffness values for a wide range of biological phenomena without the confounding effects of durotaxis. This system has the potential for wide adoption in the cell biology community because of its ease of fabrication, simple material ingredients, and wide gradient possibilities in a single well. The spatial presentation of mechanical information is a key parameter for cell behavior. We have developed a method of polymerization control in which the differential diffusion distance of unreacted cross-linker and monomer into a prepolymerized hydrogel sink results in a tunable stiffness gradient at the cell–matrix interface. This simple, low-cost, robust method was used to produce polyacrylamide hydrogels with stiffness gradients of 0.5, 1.7, 2.9, 4.5, 6.8, and 8.2 kPa/mm, spanning the in vivo physiological and pathological mechanical landscape. Importantly, three of these gradients were found to be nondurotactic for human adipose-derived stem cells (hASCs), allowing the presentation of a continuous range of stiffnesses in a single well without the confounding effect of differential cell migration. Using these nondurotactic gradient gels, stiffness-dependent hASC morphology, migration, and differentiation were studied. Finally, the mechanosensitive proteins YAP, Lamin A/C, Lamin B, MRTF-A, and MRTF-B were analyzed on these gradients, providing higher-resolution data on stiffness-dependent expression and localization.

A Particle Multi-Target Tracker for Superpositional Measurements Using Labeled Random Finite Sets

In this paper we present a general solution for multi-target tracking with superpositional measurements. Measurements that are functions of the sum of the contributions of the targets present in the surveillance area are called superpositional measurements. We base our modelling on Labeled Random Finite Set (RFS) in order to jointly estimate the number of targets and their trajectories. This modelling leads to a labeled version of Mahlers multi-target Bayes filter. However, a straightforward implementation of this tracker using Sequential Monte Carlo (SMC) methods is not feasible due to the difficulties of sampling in high dimensional spaces. We propose an efficient multi-target sampling strategy based on Superpositional Approximate CPHD (SA-CPHD) filter and the recently introduced Labeled Multi-Bernoulli (LMB) and Vo-Vo densities. The applicability of the proposed approach is verified through simulation in a challenging radar application with closely spaced targets and low signal-to-noise ratio.

Joint Initialization and Tracking of Multiple Moving Objects Using Doppler Information

In this correspondence, a new multi-target tracking (MTT) algorithm based on the probability hypothesis density (PHD) filtering framework is designed in order to improve tracking performance via the proposal of two contributions. First, unlike typical existing systems, Doppler information is additionally employed to enhance the clutter rejection capability. Specifically, position and Doppler measurements are iteratively incorporated in a two-step process based on a Gaussian mixture PHD (GMPHD) filter. Second, a concrete initialization process is proposed in the birth intensity design of the GMPHD. The initialization process from consecutive measurements leads to a reliable birth intensity that improves track management performance. Both contributions are subsequently evaluated through MTT simulations, the results of which verify that the proposed algorithm is viable.

Data fusion of radar and image measurements for multi-object tracking via Kalman filtering

Data fusion is an important issue for object tracking in autonomous systems such as robotics and surveillance. In this paper, we present a multiple-object tracking system whose design is based on multiple Kalman filters dealing with observations from two different kinds of physical sensors. Hardware integration which combines a cheap radar module and a CCD camera has been developed and data fusion method has been proposed to process measurements from those modules for multi-object tracking. Due to the limited resolution of bearing angle measurements of the cheap radar module, CCD measurements are used to compensate for the low angle resolution. Conversely, the radar module provides radial distance information which cannot be measured easily by the CCD camera. The proposed data fusion enables the tracker to efficiently utilize the radial measurements of objects from the cheap radar module and 2D location measurements of objects in image space of the CCD camera. To achieve the multi-object tracking we combine the proposed data fusion method with the integrated probability data association (IPDA) technique underlying the multiple-Kalman filter framework. The proposed complementary system based on the radar and CCD camera is experimentally evaluated through a multi-person tracking scenario. The experimental results demonstrate that the implemented system with fused observations considerably enhances tracking performance over a single sensor system.

Fast communication: Efficient importance sampling function design for sequential Monte Carlo PHD filter

In this paper, we propose a novel implementation of the probability hypothesis density (PHD) filter based on the sequential Monte Carlo (SMC) method called SMC-PHD filter. The SMC-PHD filter is analogous to the sequential importance sampling which generates samples using an importance sampling (IS) function. Even though this filter permits general class of IS density function, many previous implementations have simply used the state transition density function. However, this approach leads to a degeneracy problem and renders the filter inefficient. Thus, we propose a novel IS function for the SMC-PHD filter using a combination of an unscented information filter and a gating technique. Further, we use measurement-driven birth target intensities because they are more efficient and accurate than selecting birth targets selected using arbitrary or expected mean target states. The performance of the SMC-PHD filter with the proposed IS function was subsequently evaluated through a simulation and it was shown to outperform the standard SMC-PHD filter and recently proposed auxiliary PHD filter.

Multi-Bernoulli filter for target tracking with multi-static Doppler only measurement

Multi-static Doppler-shift has re-emerged recently in the target tracking literature along with passive sensing, especially for aircraft tracking. Tracking with multi-static Doppler only measurement requires efficient multi-sensor fusion approach and optimal sensor network configuration if possible. In this paper, we present a solution for multi-target tracking with Doppler only measurements using the multi-Bernoulli filter. To utilize Doppler measurements from multiple sensors, we investigate different multi-sensor fusion schemes and the sensor-target geometry analysis for optimal multi-static Doppler sensor network configuration. Sensor-target geometry analysis is presented to investigate optimal multi-static Doppler sensor network configuration. Numerical results verify that the proposed sequential Monte Carlo (SMC) multi-Bernoulli filter with sequential update scheme and using the carefully chosen network shows good performance. HighlightsThe multi-sensor multi-Bernoulli filter in different fusion schemes.Sensor-target geometry analysis of multi-static Doppler measurement for optimal network configuration.An efficient solution to tracking multiple targets from multi-static Doppler measurement by the multi-Bernoulli filter.

Spatio-Temporal Auxiliary Particle Filtering With

In this paper, we propose an efficient and accurate visual tracker equipped with a new particle filtering algorithm and robust subspace learning-based appearance model. The proposed visual tracker avoids drifting problems caused by abrupt motion changes and severe appearance variations that are well-known difficulties in visual tracking. The proposed algorithm is based on a type of auxiliary particle filtering that uses a spatio-temporal sliding window. Compared to conventional particle filtering algorithms, spatio-temporal auxiliary particle filtering is computationally efficient and successfully implemented in visual tracking. In addition, a real-time robust principal component pursuit (RRPCP) equipped with l1-norm optimization has been utilized to obtain a new appearance model learning block for reliable visual tracking especially for occlusions in object appearance. The overall tracking framework based on the dual ideas is robust against occlusions and out-of-plane motions because of the proposed spatio-temporal filtering and recursive form of RRPCP. The designed tracker has been evaluated using challenging video sequences, and the results confirm the advantage of using this tracker.

\ell_{1}

A receding horizon filtering problem for linear dynamic systems is considered. In receding horizon strategy, past measurements outside horizon window are discarded. In this case horizon initial state has an arbitrary mean and infinite covariance. Under these conditions we present an optimal horizon finite-impulse response filter for continuous-time linear systems. The filter is described by differential equations with initial conditions determined by the Lyapunov equations

-Norm-Based Appearance Model Learning for Robust Visual Tracking

In this paper, we propose a new method for 3D people tracking with RGB-D observations. The proposed method fuses RGB and depth data via a switching observation model. Specifically, the proposed switching observation model intelligently exploits both final detection results and raw signal intensity in a complementary manner in order to cope with missing detections. In real-world applications, the detector response to RGB data is frequently missing. When this occurs the proposed algorithm exploits the raw depth signal intensity. The fusion of detection result and raw signal intensity is integrated with the tracking task in a principled manner via the Bayesian paradigm and labeled random finite set (RFS). Our case study shows that the proposed method can reliably track people in a recently published 3D indoor data set.