Survey of visual recognition with deep learning on new domains using only a few examples

Get absolute depth in monocular self-supervised depth estimation by global scale factor with Dana

3D object detection boxes directly from image and point data using multi-modal features with CMT

Segmentation with only a few examples by using image captions instead of pixel labels with IMR-HSNet

Segment unknown objects using top-down learning and bottom-up segmentations with UDOS

Survey of semantic segmentation for autonomous driving including efficiency and use of depth or time

Detect objects in new domain without labels in target by using domain-invariant frequencies with FIT

Count objects in image with zero examples using patches like a generated sample with ZSC

Small object detection using bounding boxes guided by confidences with C-BBL

Object segmentation with only image labels using intermediate patch features with ToCo

Indoor 3D scene reconstruction using plane-regularized signed distance field with P2SDF

Complete point clouds by using missing part sensitive transformer with ProxyFormer

Survey of action labeling over time in video

Complete 3D scene using only images having occlusions by masked autoencoder with VoxFormer

3D semantic segmentation using RGB and depth with PDCNet

Get 3D human in video by self-supervised scene decomposition without prior datasets with Vid2Avatar

Reconstruct 3D object from one image using diffusion image generator with RealFusion

Survey of semi-supervised semantic segmentation

Survey of machine learning on the edge

Monocular depth using transformer, CNN, and uncertainty with URCDC-Depth