C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection

By Fang Wan, Chang Liu, Wei Ke, Xiangyang Ji, Jianbin Jiao, Qixiang Ye. 2019

258 citations recorded by the source corpus.

Weakly supervised object detection (WSOD) is a challenging task when provided with image category supervision but required to simultaneously learn object locations and object detectors. Many WSOD approaches adopt multiple instance learning (MIL) and have non-convex loss functions which are prone to get stuck into local minima (falsely localize object parts) while missing full object extent during training. In this paper, we introduce a continuation optimization method into MIL and thereby creating continuation multiple instance learning (C-MIL), with the intention of alleviating the non-convexity problem in a systematic way. We partition instances into spatially related and class related subsets, and approximate the original loss function with a series of smoothed loss functions defined within the subsets. Optimizing smoothed loss functions prevents the training procedure falling prematurely into local minima and facilitates the discovery of Stable Semantic Extremal Regions (SSERs) which indicate full object extent. On the PASCAL VOC 2007 and 2012 datasets, C-MIL improves the state-of-the-art of weakly supervised object detection and weakly supervised object localization with large margins.

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

Attention on Attention for Image Captioning (2019, methodology): “Earlier approaches to image captioning are rule/template-based [48, 34] which generate slotted caption templates and use the outputs of object detection [30, 39, 38], attribute prediction and scene recognition to fill in the slots.”
Open-Vocabulary Object Detection Using Captions (2020, background): “This problem is typically addressed using multiple instance learning, which is a well-studied topic [4, 39, 8].”
A Survey of Deep Learning-Based Object Detection (2019, background): “[217] propose continuation multiple instance learning to alleviate the non-convexity problem in”
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].”
Weakly Supervised Object Localization and Detection: A Survey (2021, unknown): “Wan-CVPR-2019 [157] Fast RCNN CNN None ImageNet(tag label) Two-stream CNN continuation instance selection and detector estimation General objects”
Mixed supervision for surface-defect detection: from weakly to fully supervised learning (2021, unknown): “Weakly supervised learning Various weakly supervised deep-learning approaches have been developed in the context of semantic segmentation and object detection [26, 29, 2, 16, 37, 36].”
Dual-stream Multiple Instance Learning Network for Whole Slide Image Classification with Self-supervised Contrastive Learning (2020, background): “In addition, deep MIL models have been considered for other weakly supervised vision tasks, including weakly supervised object localization [8] and detection [42, 46].”
Detecting Twenty-thousand Classes using Image-level Supervision (2022, background): “Many works use only image-level labels without any box supervision [26, 46, 47, 57, 64].”
Prototype Mixture Models for Few-shot Semantic Segmentation (2020, background): “However, annotating object masks for large-scale datasets is laborious, expensive, and can be impractical [9,10,11].”
LayerCAM: Exploring Hierarchical Class Activation Maps for Localization (2021, background): “Some researchers [47], [48], [49], [50], [51], [52] attempt to solve WSOL as a multiple instance learning framework.”
Comprehensive Attention Self-Distillation for Weakly-Supervised Object Detection (2020, methodology): “6 C-MIL [13] 50.”
Learning to Match Anchors for Visual Object Detection (2021, background): “To represent objects of various appearance, aspect ratios and poses with limited convolutional features, many CNN-based detectors leverage anchor boxes as reference points for object localization [1]–[7], [9], [10].”
C-MIDN: Coupled Multiple Instance Detection Network With Segmentation Guidance for Weakly Supervised Object Detection (2019, unknown): “Multiple Instance Detection Network In recent years, many end-to-end frameworks have been proposed to combine MIL and CNN [4, 31, 30, 15, 42, 36, 35, 34, 37, 25, 32, 9].”