Seeing Beyond the Visible
8 papers with code • 2 benchmarks • 2 datasets
The objective of this challenge is to automate the process of estimating the soil parameters, specifically, potassium (KKK), phosphorus pentoxide (P2O5P_2O_5P2O5), magnesium (MgMgMg) and pHpHpH, through extracting them from the airborne hyperspectral images captured over agricultural areas in Poland (the exact locations are not revealed). To make the solution applicable in real-life use cases, all the parameters should be estimated as precisely as possible.
KITTI360-EX
HYPERVIEW
Most implemented papers
Resolution-robust Large Mask Inpainting with Fourier Convolutions
saic-mdal/lama
•
•
We find that one of the main reasons for that is the lack of an effective receptive field in both the inpainting network and the loss function.
Beyond the Field-of-View: Enhancing Scene Visibility and Perception with Clip-Recurrent Transformer
In this paper, we propose the concept of online video inpainting for autonomous vehicles to expand the field of view, thereby enhancing scene visibility, perception, and system safety.
Wide-Context Semantic Image Extrapolation
shepnerd/outpainting_srn
•
•
CVPR 2019
This paper studies the fundamental problem of extrapolating visual context using deep generative models, i. e., extending image borders with plausible structure and details.
Learning Joint Spatial-Temporal Transformations for Video Inpainting
In this paper, we propose to learn a joint Spatial-Temporal Transformer Network (STTN) for video inpainting.
Towards An End-to-End Framework for Flow-Guided Video Inpainting
Optical flow, which captures motion information across frames, is exploited in recent video inpainting methods through propagating pixels along its trajectories.
FuseFormer: Fusing Fine-Grained Information in Transformers for Video Inpainting
ruiliu-ai/fuseformer
•
•
ICCV 2021
On the contrary, the soft composition operates by stitching different patches into a whole feature map where pixels in overlapping regions are summed up.
Reduce Information Loss in Transformers for Pluralistic Image Inpainting
liuqk3/put
•
•
CVPR 2022
The indices of quantized pixels are used as tokens for the inputs and prediction targets of transformer.
Predicting Soil Properties from Hyperspectral Satellite Images
ridvansalihkuzu/hyperview_eagleeyes
•
•
Conference 2022
The AI4EO HYPERVIEW challenge seeks machine learning methods that predict agriculturally relevant soil parameters (K, Mg, P2O5, pH) from airborne hyperspectral images.
