@@ -0,0 +1,13 @@
+# AttentionViG blog figures
+
+Assets for the LoG 2026 AttentionViG blog post.
+
+Source: Hakan Emre Gedik et al., [AttentionViG: Cross-Attention-Based Dynamic Neighbor Aggregation in Vision GNNs](https://arxiv.org/abs/2509.25570).
+
+- `01_cross_attention.png`: cross-attention aggregation diagram from the paper.
+- `02_architecture.png`: multi-scale architecture diagram from the paper.
+- `03_imagenet_results.png`: visual summary of the paper's ImageNet-1K results, with the knowledge-distillation setting noted.
+- `04_dense_prediction.png`: visual summary of the paper's COCO and ADE20K results.
+- `05_neighbor_visualization.png`: query-key similarity visualizations from the paper.
+
+The performance summaries are presentation graphics, not reproductions of the paper's tables.