computer-vision

Active Object Detection

3260 papers • 126 benchmarks • 313 datasets

Active Learning for Object Detection

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in active-object-detection

Trend

Dataset

Best Model

Actions

PASCAL VOC 07+12

MS COCO

Libraries

i

Use these libraries to find active-object-detection models and implementations

Datasets

No datasets available.

Subtasks

No subtasks available.

Most implemented papers

The MECCANO Dataset: Understanding Human-Object Interactions from Egocentric Videos in an Industrial-like Domain

Antonino Furnari, G. Farinella, Francesco Ragusa, S. Livatino•Sun Oct 11 2020

The MECCANO dataset is introduced, the first dataset of egocentric videos to study human-object interactions in industrial-like settings and is a revisited version of the standard human- object interaction detection task.

128

Content

0

Paper Graph

Sequential Voting with Relational Box Fields for Active Object Detection

Kris M. Kitani, Qichen Fu, Xingyu Liu•Wed Oct 20 2021

This work adopts standard RL to learn the voting function parameters and shows that it provides a meaningful improvement over a standard supervised learning approach, and performs experiments on two large-scale datasets: 100DOH and MECCANO, improving AP50 performance by 8% and 30% over the state of the art.

14 0

Paper Graph

MUS-CDB: Mixed Uncertainty Sampling With Class Distribution Balancing for Active Annotation in Aerial Object Detection

Dong Liang, Jingwei Zhang, Ying-Peng Tang, Sheng-Jun Huang•Mon Dec 05 2022

Both object-level and image-level informativeness are considered in the object selection to refrain from redundant and myopic querying and an easy-to-use class-balancing criterion is incorporated to favor the minority objects to alleviate the long-tailed class distribution problem in model training.

21 0

Paper Graph

Multiple Instance Active Learning for Object Detection

Xiangyang Ji, Fang Wan, Qixiang Ye, Tianning Yuan, Mengying Fu, Jianzhuang Liu, Songcen Xu•Mon Apr 05 2021

MI-AOD defines an instance uncertainty learning module, which leverages the discrepancy of two adversarial instance classifiers trained on the labeled set to predict instance uncertainty of the unlabeled set, and estimates the image uncertainty by re-weighting instances in a multiple instance learning (MIL) fashion.

149 0

Paper Graph

Adding a benchmark result helps the community track progress.