Global optimization for coupled detection and data association in multiple object tracking

Abstract

We present a novel framework for tracking multiple objects imaged from one or more static cameras, where the problems of object detection and data association are expressed by a single objective function. Particularly, we combine a sparsity-driven detector with the network-flow data association technique. The framework follows the Lagrange dual decomposition strategy, taking advantage of the often complementary nature of the two subproblems. Our coupling formulation avoids the problem of error propagation from which traditional "detection-tracking approaches" to multiple object tracking suffer. We also eschew common heuristics such as "non-maximum suppression" of hypotheses by modeling the joint image likelihood as opposed to applying independent likelihood assumptions. Our coupling algorithm is guaranteed to converge and can resolve the ambiguities in track maintenance due to frequent occlusion and indistinguishable appearance between objects. Furthermore, our method does not have severe scalability issues but can process hundreds of frames at the same time. Our experiments involve challenging, notably distinct datasets and demonstrate that our method can achieve results comparable to or better than those of state-of-art approaches.

SpaRTA Tracking Across Occlusions via Partitioning of 3D Clouds of Points

“…The performance of SpaRTA is reported in Table 1 and compared with four other methods: MHT [9], SDD-MHT [9], CP(LDQD) [2] and GReTA [11]. The difference between the two datasets is that one of them is sparse, and the data sequence is rather long (1100 frames), while the second dataset is dense, with a far shorter sequence (200 frames), see Fig.6.…”

“…T RACKING large groups of targets in 3D space is a challenging topic, which is particularly relevant in the field of turbulence [1], collective animal behavior [2], [3] and social sciences [4], [5] as well as in robotics [6] and autonomous mobility [7]. The technological progress of the last decades gave a boost to the development of new experimental strategies to collect 3D data, such as RGB-D, multicamera, lidar and radar systems.…”

An effective and efficient approximate two-dimensional dynamic programming algorithm for supporting advanced computer vision applications

“…Another classical example is that of sequence alignment. Generally-speaking, computer vision applications are emerging trends for such a context, and, recently, the research community has devoted a lot of attention to this topic (e.g., [2,3,4,5,6,7,8,9,10,11]). DP has been applied to various tasks in pattern recognition and computer vision [12,13].…”

Multiple People Tracking Using Camera Networks with Overlapping Views

“…Multiple tracking hypotheses are kept and the most likely trajectories are found by searching forward as well as backwards in time. The authors in [9] expressed the problems of detection and tracking in optimizing a single objective function and solved the optimization problem using Lagrange dual decomposition strategy. Specifically, the subproblem of detection is solved by using spars recovery technique and subproblem of data association is solved by network flow and the two subproblems reach consensus iteratively by projected subgradient optimization.…”