AI Paper Atlas

Research map / Detection and segmentation / Open-set and few-shot detection

Learning Object Placement by Inpainting for Compositional Data Augmentation

Lingzhi Zhang, Tarmily Wen, Jie Min, Jiancong Wang, David Han, Jianbo Shi

ECCV 2020 · Detection and segmentation › Open-set and few-shot detection · approach: Object-level representation learning · 71 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add diversity loss to the cVAE-GAN placement model (cVAE-GAN+Div)cVAE-GAN · Object placement diversitydiversity: 0.0219 -> 0.0335Improved
use PlaceNet pairwise-distance diversity loss with spectral-norm discriminator instead of cVAE-GAN+Divplacement model · Object placementFID: 53.54 -> 48.15Improved
use PlaceNet placements instead of random placement for augmentationYOLOv3 on KITTI · KITTI object detectionoverall mAP: 0.203 -> 0.371Improved
augment training data with random placementYOLOv3 on KITTI · KITTI object detectionoverall mAP: 0.359 -> 0.203Hurt
augment with PlaceNet object placementYOLOv3 · Cityscapes object detectionoverall mAP: 0.265 -> 0.302Improved
augment with PlaceNet object placementYOLOv3 · KITTI object detectionoverall mAP: 0.359 -> 0.371Improved
augment with PlaceNet object placementYOLOv3 · Cityscapes object detection, car classmAP: 0.754 -> 0.749Hurt
augment with PlaceNet object placementYOLOv3 · Cityscapes object detection, rider classmAP: 0.238 -> 0.381Improved
augment with PlaceNet object placementMask R-CNN · Cityscapes instance segmentationoverall mAP: 0.286 -> 0.302Improved
augment with PlaceNet object placementMask R-CNN · KITTI instance segmentationoverall mAP: 0.235 -> 0.307Improved
augment with random placementMask R-CNN · Cityscapes instance segmentationoverall mAP: 0.286 -> 0.279Hurt
use diverse PlaceNet placements instead of less diverse cVAE-GANYOLOv3 on KITTI · KITTI object detectionoverall mAP: 0.274 -> 0.371Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy