Data reigns supreme 🥇

Every day it becomes more evident that data is the limiting factor for state-of-the-art 📈 machine learning. Your model architecture may be revolutionary, but without high-quality data 📊 to train on, it will be doomed to mediocrity.

Pair idea with execution and use top-notch data in your next project!

We've combed through the 2359 papers accepted to CVPR in 2023 and compiled a short-list of papers introducing exciting new datasets.

Title	Tags	Paper	Dataset	Code
MVImgNet: A Large-scale Dataset of Multi-view Images	multi-view, image
GeoNet: Benchmarking Unsupervised Adaptation across Geographies	geolocation, image
Joint HDR Denoising and Fusion: A Real-World Mobile HDR Image Dataset	denoising, image
Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo	optical flow, stereo, image
ImageNet-E: Benchmarking Neural Network Robustness via Attribute Editing	image, editing
ARKitTrack: A New Diverse Dataset for Tracking Using Mobile RGB-D Data	RGB-D, segmentation, video
Diverse Embedding Expansion Network and Low-Light Cross-Modality Benchmark for Visible-Infrared Person Re-identification	low-light, cross-modal, IR
JRDB-Pose: A Large-scale Dataset for Multi-Person Pose Estimation and Tracking	pose estimation, image, keypoint, tracking
A New Benchmark: On the Utility of Synthetic Data with Blender for Bare Supervised Learning and Downstream Domain Adaptation	synthetic, domain adaptation, supervised

We would love your help in making this repository even better! If we missed a paper that introduced a new dataset, or if you can think of any ways to improve the repository, feel free to open an issue or a pull request.

This repository is inspired by paperswithcode, and the template was adapted from top-cvpr-2023-papers.