A Review and Efficient Implementation of Scene Graph Generation Metrics
This work addresses a gap in evaluation standards for scene graph generation researchers, but it is incremental as it focuses on clarifying and implementing existing metrics rather than proposing new ones.
The paper tackles the lack of precise definitions for metrics in scene graph generation by providing a review and clear definitions of commonly used metrics, and introduces a Python package and web service to facilitate their usage and benchmarking.
Scene graph generation has emerged as a prominent research field in computer vision, witnessing significant advancements in the recent years. However, despite these strides, precise and thorough definitions for the metrics used to evaluate scene graph generation models are lacking. In this paper, we address this gap in the literature by providing a review and precise definition of commonly used metrics in scene graph generation. Our comprehensive examination clarifies the underlying principles of these metrics and can serve as a reference or introduction to scene graph metrics. Furthermore, to facilitate the usage of these metrics, we introduce a standalone Python package called SGBench that efficiently implements all defined metrics, ensuring their accessibility to the research community. Additionally, we present a scene graph benchmarking web service, that enables researchers to compare scene graph generation methods and increase visibility of new methods in a central place. All of our code can be found at https://lorjul.github.io/sgbench/.