Add MapAnything to website

Author: Nik-V9

_posts/2025-09-17-mapanything.md

@@ -0,0 +1,23 @@
+---
+layout: post
+title: "MapAnything: Universal Feed-Forward Metric 3D Reconstruction"
+date: 2025-09-17 11:00:00
+categories: research
+description: "MapAnything is a simple, end-to-end trained transformer model that directly regresses the factored metric 3D geometry of a scene given various types of inputs (images, calibration, poses, or depth)."
+author: "Nikhil Keetha"
+published: true
+redirect: "https://map-anything.github.io/"
+show_sidebar: false
+# slim_content_width: true
+permalink: /mapanything_redirect/
+image: /img/posts/2025-09-17-mapanything/mapanything_teaser.png
+datatable: true
+title_image: None
+hero_image: /img/posts/2025-09-17-mapanything/mapanything_teaser.png
+hero_height: is-large
+remove_hero_title: false
+menubar_toc: false
+tags: Perception, Learning, SLAM
+---
+
+MapAnything is a simple, end-to-end trained transformer model that directly regresses the factored metric 3D geometry of a scene given various types of inputs (images, calibration, poses, or depth).