We Don’t Need No Bounding-Boxes: Training Object Class Detectors Using Only Human Verification

By Dim P. Papadopoulos, Jasper R. R. Uijlings, Frank Keller, Vittorio Ferrari. 2016

144 citations recorded by the source corpus.

Training object class detectors typically requires a large set of images in which objects are annotated by boundingboxes. However, manually drawing bounding-boxes is very time consuming. We propose a new scheme for training object detectors which only requires annotators to verify bounding-boxes produced automatically by the learning algorithm. Our scheme iterates between re-training the detector, re-localizing objects in the training images, and human verification. We use the verification signal both to improve re-training and to reduce the search space for re-localisation, which makes these steps different to what is normally done in a weakly supervised setting. Extensive experiments on PASCAL VOC 2007 show that (1) using human verification to update detectors and reduce the search space leads to the rapid production of high-quality bounding-box annotations, (2) our scheme delivers detectors performing almost as good as those trained in a fully supervised setting, without ever drawing any bounding-box, (3) as the verification task is very quick, our scheme substantially reduces total annotation time by a factor 6×-9×.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Evaluation & Benchmarks

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

In Computer Vision and Pattern Recognition

In Evaluation & Benchmarks

Citations in the corpus

Extreme Clicking for Efficient Object Annotation (2017, methodology): “We also compare to two methods which strike a trade-off between accuracy and efficiency [54, 47].”
nocaps: novel object captioning at scale (2019, neutral): “As with previous work, this task setting is also motivated by the observation that collecting human-annotated captions is resource intensive and scales poorly as object diversity grows, while on the other hand, large-scale object classification and detection datasets already exist [8,18] and their c”
The Open Images Dataset V4 (2018, methodology): “About 10 % of the bounding boxes in the training set were annotated using box veriﬁcation series, in which annotators verify bounding boxes produced automatically by a learning algorithm (Papadopoulos et al., 2016).”
Interactive Full Image Segmentation by Considering All Regions Jointly (2018, methodology): “The work of [42] annotates bounding boxes using only human verification, while [29] trained agents to determine whether it is more efficient to verify or draw a bounding box.”
Y Qualification Main task Y Return N N Y User qualified ? N Feedback Pass ? Pass ? Annotator training Annotating images (2017, methodology): “We also compare to two methods which strike a trade-off between accuracy and efficiency [54, 47].”
Object Detection in 20 Years: A Survey (2019, unknown): “Multi-instance learning is a group of supervised learning algorithms that has seen widespread application in WSOD [202], [203], [204], [205], [206], [207], [208], [209].”
Continuous Adaptation for Interactive Object Segmentation by Learning from Corrections (2019, background): “For example, when supervision for the first frame of a test video is available [43,55,13], or after annotating a subset of an image dataset [42,61].”