Multiple Instance Detection Network with Online Instance Classifier Refinement

By Peng Tang, Xinggang Wang, Xiang Bai. 2017

481 citations recorded by the source corpus.

Of late, weakly supervised object detection is with great importance in object recognition. Based on deep learning, weakly supervised detectors have achieved many promising results. However, compared with fully supervised detection, it is more challenging to train deep network based detectors in a weakly supervised manner. Here we formulate weakly supervised detection as a Multiple Instance Learning (MIL) problem, where instance classifiers (object detectors) are put into the network as hidden nodes. We propose a novel online instance classifier refinement algorithm to integrate MIL and the instance classifier refinement procedure into a single deep network, and train the network end-to-end with only image-level supervision, i.e., without object location information. More precisely, instance labels inferred from weak supervision are propagated to their spatially overlapped instances to refine instance classifier online. The iterative instance classifier refinement procedure is implemented using multiple streams in deep network, where each stream supervises its latter stream. Weakly supervised object detection experiments are carried out on the challenging PASCAL VOC 2007 and 2012 benchmarks. We obtain 47% mAP on VOC 2007 that significantly outperforms the previous state-of-the-art.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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

Citations in the corpus

Learning Open-World Object Proposals without Learning to Classify (2021, background): “Weakly-supervised approaches [5, 7, 17, 54, 62] aim to utilize abundant image-level labels and leverage classagnostic box proposals to build detectors.”
Seeing Out of tHe bOx: End-to-End Pre-training for Vision-Language Representation Learning (2021, background): “Although some works try to train detection model[38, 46] with weakly-supervised, the performance is far below the requirements.”
Single-Shot Object Detection with Enriched Semantics (2017, background): “With the emergence of deep neural networks, computer vision has been improved significantly in many aspects such as image classification [11, 13, 15, 24, 27], object detection [3, 17, 20, 21, 25], and segmentation [2, 10, 18].”
Open-Vocabulary Instance Segmentation via Robust Cross-Modal Pseudo-Labeling (2021, background): “Following recent works on weakly-supervised learning [96, 97], we select the highest confident bounding box for each object to minimize false-positive predictions.”
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”
Collaborative Learning for Weakly Supervised Object Detection (2018, methodology): “The first includes the MIL methods [Cinbis et al., 2017; Wang et al., 2014] and various end-to-end MIL-CNN models [Bilen and Vedaldi, 2016; Kantorov et al., 2016; Tang et al., 2017] following the twostream structure of WSDDN [Bilen and Vedaldi, 2016].”
Angular Visual Hardness (2019, background): “Measuring sample conﬁdence is also important in partially-supervised problems such as semi-supervised learning [71, 72], unsupervised domain adaptation [8] and weakly-supervised learning [65] due to their under-constrained nature.”
C-MIDN: Coupled Multiple Instance Detection Network With Segmentation Guidance for Weakly Supervised Object Detection (2019, methodology): “[31] combine WSDDN with several instance classiﬁers, and pro-pose an online instance classiﬁer reﬁnement method to re-ﬁne the initial candidates of WSDDN. PCL [30] uses a graph-based center cluster method and average MIL loss based on [31].”
From Discriminant to Complete: Reinforcement Searching-Agent Learning for Weakly Supervised Object Detection (2020, methodology): “[9] design an online instance classifier refinement (OICR) approach and embed it into the WSDDN via OICR layers.”
Multi-Scale Positive Sample Refinement for Few-Shot Object Detection (2020, background): “Research basically concentrates on how to rank and classify region proposals with only coarse labels through multiple instance learning [36,37,39].”
C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection (2019, methodology): “With VGGF, C-MIL respectively outperformed the WCCN [14], OICR [34], and MELM [37] We further re-trained an Fast-RCNN detector using the learned pseudo objects as ground-truth, and achieved 53.1% mAP, as shown in Table 3, which outperformed the state-of-the-art methods by 2.7% ∼ 6.1%.”
SIXray: A Large-Scale Security Inspection X-Ray Benchmark for Prohibited Item Discovery in Overlapping Images (2019, background): “This is related to the research in weakly-supervised object localization [3] [6] [33], or a series of work in localizing objects using top-down class activation [8] [7] [40].”
Detecting 11K Classes: Large Scale Object Detection Without Fine-Grained Bounding Boxes (2019, background): “OICR [30] improves WSDDN by incorporating multiple reﬁnement streams with pseudo ground-truth.”
Cross-Domain Weakly-Supervised Object Detection Through Progressive Domain Adaptation (2018, methodology): “For Faster R-CNN [28], we used the reimplementation provided in ChainerCV [25].”
Weakly-Supervised Semantic Segmentation Network with Deep Seeded Region Growing (2018, background): “Therefore, a promising direction in the computer vision research is to develop object recognition methods that can learn from unlabeled or weakly labeled images [14, 32].”
PCL: Proposal Cluster Learning for Weakly Supervised Object Detection (2018, methodology): “This paper is an extended version of our previous work [21].”
SuperYOLO: Super Resolution Assisted Object Detection in Multimodal Remote Sensing Imagery (2022, background): “DigitalObjectIdentiﬁer 10.1109/TGRS.2023.3258666 neural network (DNN)-based object detection frameworks [1], [2], [3], [4], [5] have been proposed, updated, and optimized in computer vision.”