DOCK: Detecting Objects by Transferring Common-Sense Knowledge

By Krishna Kumar Singh, Santosh Kumar Divvala, Ali Farhadi, Yong Jae Lee. 2018

31 citations recorded by the source corpus.

We present a scalable approach for Detecting Objects by transferring Common-sense Knowledge (DOCK) from source to target categories. In our setting, the training data for the source categories have bounding box annotations, while those for the target categories only have image-level annotations. Current state-of-the-art approaches focus on image-level visual or semantic similarity to adapt a detector trained on the source categories to the new target categories. In contrast, our key idea is to (i) use similarity not at the image-level, but rather at the region-level, and (ii) leverage richer common-sense (based on attribute, spatial, etc.) to guide the algorithm towards learning the correct detections. We acquire such common-sense cues automatically from readily-available knowledge bases without any extra human effort. On the challenging MS COCO dataset, we find that common-sense knowledge can substantially improve detection performance over existing transfer-learning baselines.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Knowledge Graphs & Representation

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Anelia Angelova
2. Weicheng Kuo
3. Tsung-Yi Lin
4. Shih-Fu Chang
5. In-So Kweon
6. Dahun Kim
7. Hao Yang
8. Hao Chen

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Dacheng Tao
3. Anelia Angelova
4. In-So Kweon
5. Jitendra Malik
6. Trevor Darrell
7. Xiaogang Wang
8. Ming-Hsuan Yang

Citations in the corpus

• Learning Open-World Object Proposals without Learning to Classify (2021, methodology): “Approaches under semi-supervised setup [27,44,49,50,55,58,73] employ the weak image-level labels for novel classes as well as boxlevel labels for base classes.”
• Detecting 11K Classes: Large Scale Object Detection Without Fine-Grained Bounding Boxes (2019, result): “We have also compared our methods with semisupervised detection methods presented in [29, 32, 35], in a same-granularity random-split setting, and achieved favorable results.”
• Learning Visual Commonsense for Robust Scene Graph Generation (2020, methodology): “Few methods have modeled commonsense as a standalone module which is late-fused into the prediction of the perception model [13,32].”