Weakly Supervised Object Localization with Latent Category Learning

By Weiqiang Ren, Kaiqi Huang, Tien-Ping Tan. 2014

198 citations recorded by the source corpus.

Subfields: Aerospace Engineering, 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 Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

What's the Point: Semantic Segmentation with Point Supervision (2015, background): “fectively teach computers about new visual concepts. Algorithms trained from large-scale carefully annotated data enjoy better performance than their weakly supervised counterparts (e.g., [11] versus [38], [6] versus [25], [22] versus [26]); however, obtaining strongly supervised data is very expens”
Self-taught object localization with deep networks (2014, background): “Furthermore, recent work has shown that models and features learned using class labels can be effectively leveraged to improve other tasks, such as detection [12], weakly-supervised localization [31], and attribute classification [34].”
Weakly Supervised Object Localization with Multi-Fold Multiple Instance Learning (2015, neutral): “[53] propose a two-step method, which first groups selective search candidate windows [49] from the positive images of a class into visual clusters and then chooses the most discriminative cluster of windows.”
Unsupervised object discovery and localization in the wild: Part-based matching with bottom-up region proposals (2015, methodology): “Typically, weakly-supervised localization [9, 35, 36, 43, 45, 56] requires positive and negative image-level labels for a target object class.”
From Discriminant to Complete: Reinforcement Searching-Agent Learning for Weakly Supervised Object Detection (2020, background): “Some of these approaches [4]–[6], [8], [9], [28], [29] adopt the deep features extracted from the pretrained deep neural networks to facilitate the learning process.”
Discover and Learn New Objects from Documentaries (2017, methodology): “The timeconsuming annotation process in object detection can be sidestepped by weakly supervised learning [2, 3, 7, 9, 13, 30, 33, 36, 40, 41, 42, 43, 44].”
C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection (2019, methodology): “WSOD approaches only require image category annotations indicating the presence or absence of a category of objects in images, significantly reducing human involvement by omitting labor-intensive bounding-box annotations [6,12,27–29,40].”
Object Detection With Deep Learning: A Review (2018, unknown): “To release this burden, semantic prior [55], unsupervised object discovery [221], multiple instance learning [222] and deep neural network prediction [47] can be integrated to make best use of image-level supervision to”
Weakly Supervised Deep Detection Networks (2015, methodology): “[36], for example, uses a pre-trained CNN to describe image regions and then learn object categories as corresponding visual topics.”
Track and Transfer: Watching Videos to Simulate Strong Human Supervision for Weakly-Supervised Object Detection (2016, methodology): “We compare with state-of-the-art weakly-supervised detection methods [33, 34, 1, 37, 4] that use the same AlexNet CNN features pre-trained on ILSVRC 2012.”
Dissimilarity Coefficient Based Weakly Supervised Object Detection (2018, methodology): “Conventional methods often treat WSOD as a Multiple Instance Learning (MIL) problem [10] by representing each image as a bag of instances (that is, putative bounding boxes) [2, 6, 31, 36, 38].”
Large-Scale Weakly Supervised Object Localization via Latent Category Learning (2015, background): “In recent years, many studies have been proposed in weakly supervised localization, e.g., the exhaustive search [1], [2], [6], [24], multiple instance learning [3]–[5], [25], [26], interintra-class modeling [7]–[9], [27] and topic model [10], [11], [28]–[30].”
W2F: A Weakly-Supervised to Fully-Supervised Framework for Object Detection (2018, methodology): “label) annotations are available in the training dataset, and there are many works [1, 34, 23, 2], which can capture the instance-level annotations.”
Min-Entropy Latent Model for Weakly Supervised Object Detection (2018, methodology): “, an image was first decomposed into object proposals, with which clustering [14]–[16], latent variable learning [12]–[15], [17] or multiple instance learning [8], [10], [11], [21], [32] was used to perform proposal selection and classifier estimation.”
PCL: Proposal Cluster Learning for Weakly Supervised Object Detection (2018, background): “Early works treat CNN models pre-trained on ImageNet as off-the-shelf feature extractors [12], [13], [14], [15], [37], [38], [39], [40].”