Consistency-based Semi-supervised Learning for Object detection

By Jisoo Jeong, Seungeui Lee, Jeesoo Kim, Nojun Kwak. Neural Information Processing Systems, 2019

484 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/1466e5837bc4717e1321fe033235d6073b0fe1cd

How researchers see this paper

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

Citations in the corpus

Unbiased Teacher for Semi-Supervised Object Detection (2021, methodology): “Table 1: Experimental results on COCO-standard comparing with CSD (Jeong et al., 2019) and STAC (Sohn et al.”
A Survey on Deep Semi-Supervised Learning (2021, methodology): “In this survey, we mainly discuss the methods applied to image classification since these methods can be extended to other applications, such as [49] applied consistency training to object detection, [50] modified Semi-GANs to fit the scenario of semantic segmentation.”
SODA10M: Towards Large-Scale Object Detection Benchmark for Autonomous Driving (2021, background): “In the field of object detection, previous works focus on training detectors with a combination of labeled, weaky-labeled or unlabeled data [26, 50, 13], while recent works [28, 47] train detectors with a small set of labeled data and a larger amount of unlabeled images.”
3rd Continual Learning Workshop Challenge on Egocentric Category and Instance Level Object Understanding (2022, background): “Research does not only cover the advancement of performance under classic learning paradigms (supervised, offline training) [20, 21, 22, 23, 24, 25, 26, 27], but also encompasses areas such as weakly-supervised learning [28, 29, 30, 31], few-shot learning [32, 33, 34, 35, 36, 37, 38], and semi-super”
Semi-Supervised and Unsupervised Deep Visual Learning: A Survey (2022, methodology): “Compared to standard consistency regularization, these meth-ods especially need re-estimating the bounding box location in an augmented image, such as ﬂip the bounding box [169], or calculate the overlapped bounding boxes of two mixed images in MixUp [51].”
Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation (2022, methodology): “Two main branches of methodology are proposed to tackle the issue, namely entropy minimization [22, 53, 37, 65, 80, 51] and consistency regularization [36, 54, 58, 64, 30, 6, 5, 48, 20, 38].”
PreDet: Large-scale weakly supervised pre-training for detection (2021, methodology): “Semi-supervised learning and Self-training Semisupervised and self-training methods [50, 62, 22, 39, 29, 70, 60, 61] train a model jointly with a specific target dataset and the provided large-scale dataset, and have shown benefits for detection.”
Omni-DETR: Omni-Supervised Object Detection with Transformers (2022, unknown): “[22] leverages the consistency regularization between two different augmented views of a single image.”
ST++: Make Self-trainingWork Better for Semi-supervised Semantic Segmentation (2021, background): “It can be a fixed value [47, 48] or be scheduled during training [27].”
Semi-supervised Medical Image Segmentation through Dual-task Consistency (2020, methodology): “Jeong et al. (2019) used consistency constraints as a tool for enhancing detection performance by making full use of available unlabeled data.”
A Simple Semi-Supervised Learning Framework for Object Detection (2020, methodology): “For STAC, we follow [23] to have diﬀerent level of unlabeled sources, including VOC12 and the subset of MS-COCO data with the same classes as PASCAL VOC.”
Comprehensive Attention Self-Distillation for Weakly-Supervised Object Detection (2020, methodology): “Besides our attention self-distillation, there are other regularization strategies emerged for semi-supervised object detection and multi-label classiﬁcation such as Prediction Consistency [46] and Attention Consistency [29].”