# Atomizer-IO: Beyond Pixels, Patches and Grids

arXiv: https://arxiv.org/abs/2609.40320 · HTML: https://mlsanity.com/paper/2609.40320/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 38 of its 49 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Builds on (most cited first, 30 of 38)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space](https://mlsanity.com/paper/1706.02413/) (2017)
- [EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification](https://mlsanity.com/paper/1709.00029/) (2017)
- [Unified Perceptual Parsing for Scene Understanding](https://mlsanity.com/paper/1807.10221/) (2018)
- [KPConv: Flexible and Deformable Convolution for Point Clouds](https://mlsanity.com/paper/1904.08889/) (2019)
- [Perceiver: General Perception with Iterative Attention](https://mlsanity.com/paper/2103.03206/) (2021)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [PointCNN: Convolution On X-Transformed Points](https://mlsanity.com/paper/1801.07791/) (2018)
- [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://mlsanity.com/paper/2107.14795/) (2021)
- [RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds](https://mlsanity.com/paper/1911.11236/) (2019)
- [Deep Learning for 3D Point Clouds: A Survey](https://mlsanity.com/paper/1912.12033/) (2019)
- [Large-scale Point Cloud Semantic Segmentation with Superpoint Graphs](https://mlsanity.com/paper/1711.09869/) (2017)
- [Scale-MAE: A Scale-Aware Masked Autoencoder for Multiscale Geospatial Representation Learning](https://mlsanity.com/paper/2212.14532/) (2022)
- [Rotary Position Embedding for Vision Transformer](https://mlsanity.com/paper/2403.13298/) (2024)
- [Foundation Models for Generalist Geospatial Artificial Intelligence](https://mlsanity.com/paper/2310.18660/) (2023)
- [Prithvi-EO-2.0: A Versatile Multi-Temporal Foundation Model for Earth Observation Applications](https://mlsanity.com/paper/2412.02732/) (2024)
- [Panoptic Segmentation of Satellite Image Time Series with Convolutional Temporal Attention Networks](https://mlsanity.com/paper/2107.07933/) (2021)
- [xBD: A Dataset for Assessing Building Damage from Satellite Imagery](https://mlsanity.com/paper/1911.09296/) (2019)
- [ShellNet: Efficient Point Cloud Convolutional Neural Networks using Concentric Shells Statistics](https://mlsanity.com/paper/1908.06295/) (2019)
- [Neural Plasticity-Inspired Multimodal Foundation Model for Earth Observation](https://mlsanity.com/paper/2403.15356/) (2024)
- [TerraMind: Large-Scale Generative Multimodality for Earth Observation](https://mlsanity.com/paper/2504.11171/) (2025)
- [Lightweight, Pre-trained Transformers for Remote Sensing Timeseries](https://mlsanity.com/paper/2304.14065/) (2023)
- [PANGAEA: A Global and Inclusive Benchmark for Geospatial Foundation Models](https://mlsanity.com/paper/2412.04204/) (2024)
- [DALES: A Large-scale Aerial LiDAR Data Set for Semantic Segmentation](https://mlsanity.com/paper/2004.11985/) (2020)
- [ConvPoint: Continuous Convolutions for Point Cloud Processing](https://mlsanity.com/paper/1904.02375/) (2019)
- [Efficient 3D Semantic Segmentation with Superpoint Transformer](https://mlsanity.com/paper/2306.08045/) (2023)
- [Lightweight Temporal Self-Attention for Classifying Satellite Image Time Series](https://mlsanity.com/paper/2007.00586/) (2020)

## Abstract (arXiv)

Most vision architectures assume that observations lie on a regular grid, an effective abstraction for natural images but a restrictive one for sensing data whose channels, temporal sampling, spatial resolution, and geometry can vary. Generic set-based architectures remove the grid, but also remove useful spatial inductive biases. We introduce Atomizer-IO, an architecture that places observations first and derives structure from their physical relationships. Building on top of an atomic representation of the data, each observation is described by its measurement and acquisition metadata, while local cross-attention maps observations to anchor points that can be arbitrarily placed. We evaluate this design by progressively relaxing the grid assumption, from varying input raster configurations and incomplete channel sets to flexible output density and, ultimately, inputs without a raster grid. Atomizer-IO is competitive with flexible EO-specific architectures on most tasks, while offering post-training control over inference cost and competitive compute--performance trade-offs. The same formulation extends without architectural redesign to unordered 3D point clouds, showing that the atomic interface generalizes beyond regular raster inputs. These results suggest that pixels, patches, and grids do not need to define the interface of a sensing architecture.