Cong Yuren, Hanno Ackermann, Wentong Liao, Michael Ying Yang, Bodo Rosenhahn
Semantic image understanding is a challenging topic in computer vision. It requires to detect all objects in an image, but also to identify all the relations between them. Detected objects, their labels and the discovered relations can be used to construct a scene graph which provides an abstract semantic interpretation of an image. In previous works, relations were identified by solving an assignment problem formulated as Mixed-Integer Linear Programs. In this work, we interpret that formulation as Ordinary Differential Equation (ODE). The proposed architecture performs scene graph inference by solving a neural variant of an ODE by end-to-end learning. It achieves state-of-the-art results on all three benchmark tasks: scene graph generation (SGGen), classification (SGCls) and visual relationship detection (PredCls) on Visual Genome benchmark.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Scene Parsing | Visual Genome | Recall@20 | 21.6 | NODIS |
| Scene Parsing | Visual Genome | Recall@50 | 27.7 | NODIS |
| 2D Semantic Segmentation | Visual Genome | Recall@20 | 21.6 | NODIS |
| 2D Semantic Segmentation | Visual Genome | Recall@50 | 27.7 | NODIS |
| Scene Graph Generation | Visual Genome | Recall@20 | 21.6 | NODIS |
| Scene Graph Generation | Visual Genome | Recall@50 | 27.7 | NODIS |