• Corpus ID: 239049881

Reconstruction of Fragmented Trajectories of Collective Motion using Hadamard Deep Autoencoders

  title={Reconstruction of Fragmented Trajectories of Collective Motion using Hadamard Deep Autoencoders},
  author={Kelum Gajamannage and Yonggi Park and Randy Clinton Paffenroth and Anura P. Jayasumana},
Learning dynamics of collectively moving agents such as fish or humans is an active field in research. Due to natural phenomena such as occlusion and change of illumination, the multi-object methods tracking such dynamics might lose track of the agents where that might result fragmentation in the constructed trajectories. Here, we present an extended deep autoencoder (DA) that we train only on fully observed segments of the trajectories by defining its loss function as the Hadamard product of a… 

Figures and Tables from this paper


Reconstruction of Agents’ Corrupted Trajectories of Collective Motion Using Low-rank Matrix Completion
This work utilizes mutual interactions and dependencies between the agents to reconstruct the missing segments of the trajectories of particles’ trajectories, and utilizes a low-rank matrix completion technique to reconstruct incomplete trajectories for two representative self-propelled particle swarms.
Identifying manifolds underlying group motion in Vicsek agents
This work constructs a novel metric which is susceptible to variations in the collective motion, thus revealing distinct underlying manifolds underlying a group of evolving agents, and provides an effective model-free framework for the analysis of collective behavior across animal species.
You'll never walk alone: Modeling social behavior for multi-target tracking
A model of dynamic social behavior, inspired by models developed for crowd simulation, is introduced, trained with videos recorded from birds-eye view at busy locations, and applied as a motion model for multi-people tracking from a vehicle-mounted camera.
Eliminating the Invariance on the Loss Landscape of Linear Autoencoders
This paper analytically identifies the structure of the associated loss surface for linear autoencoders (LAEs) and establishes an analytical expression for the set of all critical points, showing that it is a subset of critical points of MSE, and that all local minima are still global.
Learning to Track and Identify Players from Broadcast Sports Videos
A system that possesses the ability to detect and track multiple players, estimates the homography between video frames and the court, and identifies the players, and proposes a novel Linear Programming (LP) Relaxation algorithm for predicting the best player identification in a video clip.
Learning affinities and dependencies for multi-target tracking using a CRF model
This work argues that the independent assumption is not valid in many cases, and adopts a CRF model to consider both tracklet affinities and dependencies among them, which are represented by unary term costs and pairwise term costs respectively.
A Nonlinear Dimensionality Reduction Framework Using Smooth Geodesics
A framework for nonlinear dimensionality reduction that generates a manifold in terms of smooth geodesics that is designed to treat problems in which manifold measurements are either sparse or corrupted by noise is proposed.
Detecting phase transitions in collective behavior using manifold's curvature
The Phase Transition Detection (PTD) method is validated using one particle simulation and three real world examples and it is attest that the same phase transition can also be approximated by singular value ratios computed locally over the data in a neighborhood on the manifold.
Model-Based Tracking of Multiple Worms and Fish
This paper addresses the problem of tracking multiple undulating creatures through interactions and partial occlusions from a top view. Detailed posture estimation of biological model organisms is
Visual tracking via adaptive structural local sparse appearance model
A simple yet robust tracking method based on the structural local sparse appearance model which exploits both partial information and spatial information of the target based on a novel alignment-pooling method and employs a template update strategy which combines incremental subspace learning and sparse representation.