-
DeepAerial
A dataset of aerial images, consisting of 500 image pairs with known affine transformations. -
SPair-71k: A large-scale benchmark for semantic correspondence
The proposed method, dubbed Dynamic Hyperpixel Flow, learns to compose hypercolumn features on the fly by selecting a small number of relevant layers from a deep convolutional... -
Learning to Compose Hypercolumns for Visual Correspondence
Feature representation plays a crucial role in visual correspondence, and recent methods for image matching resort to deeply stacked convolutional layers. -
Efficient Spatially Adaptive Convolution and Correlation
The dataset used in the paper for pattern matching and image processing tasks. -
Boreas: A multi-season autonomous driving dataset
Boreas: A multi-season autonomous driving dataset. -
ORB: An efficient alternative to SIFT or SURF
ORB: An efficient alternative to SIFT or SURF. -
Hpatches: A Benchmark and Evaluation of Handcrafted and Learned Local Descrip...
A benchmark and evaluation of handcrafted and learned local descriptors for keypoint detection task. -
Dual-Resolution Correspondence Networks
Establishing dense pixel-wise correspondences between a pair of images in a coarse-to-fine manner. -
MegaDepth dataset
The dataset used for training the D2-Net model, consisting of 327,036 image pairs with at least 50% overlap in the sparse SfM point cloud. -
VIGOR: Cross-View Image Geo-localization beyond One-to-one Retrieval
A new large-scale benchmark for cross-view image geo-localization beyond one-to-one retrieval, which is a more realistic setting for real-world applications. -
HPatches dataset
The HPatches dataset evaluates the mAP in three different tasks: patch verification, image matching and patch retrieval.
