This paper addresses a novel task of anticipating 3D human-object
intera...
LiDAR-based 3D detection plays a vital role in autonomous navigation.
Su...
Depth completion, which aims to generate high-quality dense depth maps f...
Existing approaches for semi-supervised object detection assume a fixed ...
In the context of few-shot learning, it is currently believed that a fix...
This paper aims to deal with the ignored real-world complexities in prio...
Closing the domain gap between training and deployment and incorporating...
This paper proposes NeuralEditor that enables neural radiance fields (Ne...
Object detectors often suffer from the domain gap between training (sour...
Object discovery – separating objects from the background without manual...
Predicting diverse human motions given a sequence of historical poses ha...
This work proposes an end-to-end multi-camera 3D multi-object tracking (...
Existing work on continual learning (CL) is primarily devoted to develop...
Although reinforcement learning has found widespread use in dense reward...
Lifelong learners must recognize concept vocabularies that evolve over t...
Being able to learn an effective semantic representation directly on raw...
Recently, it has been observed that a transfer learning solution might b...
Self-supervised contrastive learning is a powerful tool to learn visual
...
Comprehensive 3D scene understanding, both geometrically and semanticall...
In the real open world, data tends to follow long-tailed class distribut...
This paper studies the problem of object discovery – separating objects ...
Recent work has suggested that a good embedding is all we need to solve ...
Recently, it has been observed that a transfer learning solution might b...
In LiDAR-based 3D object detection for autonomous driving, the ratio of ...
Learning accurate classifiers for novel categories from very few example...
Few-shot classification aims at classifying categories of a novel task b...
We present a novel semi-supervised semantic segmentation method which jo...
Generative modeling has recently shown great promise in computer vision,...
Learning to detect novel objects from few annotated examples is of great...
Training on datasets with long-tailed distributions has been challenging...
This paper presents a detection-aware pre-training (DAP) approach, which...
Learning to detect an object in an image from very few training examples...
We consider the few-shot classification task with an unbalanced dataset,...
Deep learning classification models typically train poorly on classes wi...
Generative modeling has recently shown great promise in computer vision,...
Embodied perception refers to the ability of an autonomous agent to perc...
The problem of rare category recognition has received a lot of attention...
CNNs have made an undeniable impact on computer vision through the abili...
Humans can robustly learn novel visual concepts even when images undergo...
One of the key limitations of modern deep learning based approaches lies...
Humans can quickly learn new visual concepts, perhaps because they can e...