Localization in the Crowd with Topological Constraints
This work addresses the problem of accurate localization of individuals in crowded scenes, which is important for applications like crowd counting and surveillance, by improving robustness to spatial semantic errors.
This paper tackles the problem of crowd localization, which involves predicting the locations of individuals in crowded scenes. The authors introduce a topological constraint and a persistence loss based on persistent homology to reduce spatial semantic errors, particularly in cluttered regions. This approach leads to improved localization quality and outperforms previous methods on multiple public benchmarks.
We address the problem of crowd localization, i.e., the prediction of dots corresponding to people in a crowded scene. Due to various challenges, a localization method is prone to spatial semantic errors, i.e., predicting multiple dots within a same person or collapsing multiple dots in a cluttered region. We propose a topological approach targeting these semantic errors. We introduce a topological constraint that teaches the model to reason about the spatial arrangement of dots. To enforce this constraint, we define a persistence loss based on the theory of persistent homology. The loss compares the topographic landscape of the likelihood map and the topology of the ground truth. Topological reasoning improves the quality of the localization algorithm especially near cluttered regions. On multiple public benchmarks, our method outperforms previous localization methods. Additionally, we demonstrate the potential of our method in improving the performance in the crowd counting task.