VEIL: Vetting Extracted Image Labels from In-the-Wild Captions for Weakly-Supervised Object Detection
This addresses the challenge of using noisy captions for object detection in computer vision, offering a domain-specific improvement for weakly-supervised learning.
The paper tackles the problem of label noise in large-scale vision-language datasets for object detection by proposing a method to vet extracted labels from noisy captions, improving weakly-supervised object detection by 30% (from 31.2 to 40.5 mAP on PASCAL VOC) without using bounding boxes.
The use of large-scale vision-language datasets is limited for object detection due to the negative impact of label noise on localization. Prior methods have shown how such large-scale datasets can be used for pretraining, which can provide initial signal for localization, but is insufficient without clean bounding-box data for at least some categories. We propose a technique to "vet" labels extracted from noisy captions, and use them for weakly-supervised object detection (WSOD), without any bounding boxes. We analyze and annotate the types of label noise in captions in our Caption Label Noise dataset, and train a classifier that predicts if an extracted label is actually present in the image or not. Our classifier generalizes across dataset boundaries and across categories. We compare the classifier to nine baselines on five datasets, and demonstrate that it can improve WSOD without label vetting by 30% (31.2 to 40.5 mAP when evaluated on PASCAL VOC). See dataset at: https://github.com/arushirai1/CLaNDataset.