Skip to search formSkip to main contentSkip to account menu

Frustratingly Simple Few-Shot Object Detection

This work finds that fine-tuning only the last layer of existing detectors on rare classes is crucial to the few-shot object detection task, and establishes a new state of the art on the revised benchmarks.
View PDF on arXiv (opens in a new tab)

Learning Hierarchical Semantic Image Manipulation through Structured Representations

A novel hierarchical framework for semantic image manipulation that allows a user to manipulate images at object-level by adding, removing, and moving one bounding box at a time is presented.
View PDF on arXiv (opens in a new tab)

Revisiting Hierarchical Approach for Persistent Long-Term Video Prediction

This work revisits hierarchical models in video prediction, and shows that modeling structures and their dynamics in the discrete semantic structure space with a stochastic recurrent estimator leads to surprisingly successful long-term prediction.
View PDF on arXiv (opens in a new tab)

Robust Object Detection via Instance-Level Temporal Cycle Confusion

This work introduces a novel self-supervised task, instance-level temporal cycle confusion (CycConf), which operates on the region features of the object detectors, and encourages the object detector to explore invariant structures across instances under various motions, which leads to improved model robustness in unseen domains at test time.
View on IEEE (opens in a new tab)

Tracking Every Thing in the Wild

A new metric, Track Every Thing Accuracy (TETA), is introduced, breaking tracking measurement into three sub-factors: localization, association, and classification, allowing comprehensive benchmarking of tracking performance even under inaccurate classification.
View PDF on arXiv (opens in a new tab)

Dense Prediction with Attentive Feature Aggregation

Attentive Feature Aggregation is introduced to fuse different network layers with more expressive non-linear operations and improves the performance of the Deep Layer Aggregation model by nearly 6% mIoU on Cityscapes.
View on IEEE (opens in a new tab)

QDTrack: Quasi-Dense Similarity Learning for Appearance-Only Multiple Object Tracking

Quasi-Dense Similarity Learning is presented, which densely samples hundreds of object regions on a pair of images for contrastive learning and which rivals the performance of state-of-the-art tracking methods on all benchmarks and sets a new state of theart on the large-scale BDD100K MOT benchmark, while introducing negligible computational overhead to the detector.
View PDF on arXiv (opens in a new tab)

Composite Learning for Robust and Effective Dense Predictions

It is found that jointly training a dense prediction task with a self-supervised (auxiliary) task can consistently improve the performance of the target task, while eliminating the need for labeling auxiliary tasks.
View on IEEE (opens in a new tab)
By clicking accept or continuing to use the site, you agree to the terms outlined in our Privacy Policy (opens in a new tab), Terms of Service (opens in a new tab), and Dataset License (opens in a new tab)