Multilabel Conditional Random Field Classification for UAV Images

Zeggada, Abdallah; Benbraika, Souad; Melgani, Farid; Mokhtari, Zouhir

doi:10.1109/lgrs.2018.2790426

Cited by 26 publications

(23 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…E.g., the existence of ships infers to a high probable co-occurrence of the sea, while the presence of buildings is almost always accompanied by coexistence of pavement. However, the recently proposed multi-label classification methods [35,36,37,38] assumed that classes are independent and employed a set of binary classifiers [35] or a regression model [36,37,38] to infer the existence of each class separately.…”

Section: The Challenges Of Multi-label Classificationmentioning

confidence: 99%

Recurrently exploring class-wise attention in a hybrid convolutional and bidirectional LSTM network for multi-label aerial image classification

Hua

Mou

Zhu

2019

ISPRS Journal of Photogrammetry and Remote Sensing

144

103

View full text Add to dashboard Cite

Aerial image classification is of great significance in the remote sensing community, and many researches have been conducted over the past few years. Among these studies, most of them focus on categorizing an image into one semantic label, while in the real world, an aerial image is often associated with multiple labels, e.g., multiple object-level labels in our case. Besides, a comprehensive picture of present objects in a given high-resolution aerial image can provide a more in-depth understanding of the studied region. For these reasons, aerial image multi-label classification has been attracting increasing attention. However, one common limitation shared by existing methods in the community is that the co-occurrence relationship of various classes, so-called class dependency, is underexplored and leads to an inconsiderate decision. In this paper, we propose a novel end-to-end network, namely class-wise attention-based convolutional and bidirectional LSTM network (CA-Conv-BiLSTM), for this task. The proposed network consists of three indispensable components: (1) a feature extraction module, (2) a class attention learning layer, and (3) a bidirectional LSTM-based sub-network. Particularly, the feature extraction module is designed for extracting fine-grained semantic feature maps, while the class attention learning layer aims at capturing discriminative class-specific features. As the most important part, the bidirectional LSTM-based sub-network models the underlying class dependency in both directions and produce structured multiple object labels. Experimental results on UCM multi-label dataset and DFC15 multi-label dataset validate the effectiveness of our model quantitatively and qualitatively.

show abstract

Section: The Challenges Of Multi-label Classificationmentioning

confidence: 99%

Recurrently exploring class-wise attention in a hybrid convolutional and bidirectional LSTM network for multi-label aerial image classification

Hua

Mou

Zhu

2019

ISPRS Journal of Photogrammetry and Remote Sensing

144

103

View full text Add to dashboard Cite

show abstract

“…CRFs assign proper probability distribution over possible labeling and it also normalize easily to analogues of stochastic context free grammars that is used for the natural language processing and prediction. Abdallah Zeggada et al In order to study the simultaneously spatial contextual information and cross-correlation between labels the multi labeling classification method of unmanned aerial vehicle (UAV) imagery within a conditional random field (CRF) model is developed [2]. This methodology consists of two main phases.…”

Section: Conditional Random Fieldmentioning

confidence: 99%

“…Unmanned aerial vehicles (UAVs) have several pros, extremely high resolution (EHR) is captured and the multi label classification can be used for analyzing these images. Structured random fields exhibit techniques used specially in EHR images, in which a single object may be consisting of thousands of pixels which is used in combining spatially neighboring information in the classification model [2]. The multi label CRF framework for EHR UAV images is used at a tile level under a CRF perspective.…”

Section: Conditional Random Fieldmentioning

confidence: 99%

A survey on video classification using action recognition

Andrew¹,

Fiona²

2018

IJET

View full text Add to dashboard Cite

The growth in multimedia technology have resulted in producing a variety of videos every day. These videos should be classified in order to help people identify the correct video which they search for when needed. The video classification problem can be said as a probabilistic data classification problem which falls as a subcategory of the machine learning technique. Classification helps in indexing, analyzing, searching etc. A survey has been made on the present technologies that are used for video classification. Various techniques used for video classification such as Multiple Instance Learning (MIL), Conditional Random Field (CRFs) and classifying based on the action and gesture are studied.

show abstract

“…For example, one of the primal multilabel methods proposed within the RS field was presented in [39] where the authors define a multilabel support vector machine (SVM) for multilabel active learning. To simultaneously exploit the spatial-contextual information and the correlation among the labels, Zeggada et al [40] presented a conditional random field (CRF) framework for multilabel classification of images collected by unmanned aerial vehicles (UAVs).…”

Section: Introductionmentioning

confidence: 99%

Graph Relation Network: Modeling Relations Between Scenes for Multilabel Remote-Sensing Image Classification and Retrieval

Kang

Fernández-Beltrán

Hong

et al. 2021

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

Due to the proliferation of large-scale remotesensing (RS) archives with multiple annotations, multilabel RS scene classification and retrieval are becoming increasingly popular. Although some recent deep learning-based methods are able to achieve promising results in this context, the lack of research on how to learn embedding spaces under the multilabel assumption often makes these models unable to preserve complex semantic relations pervading aerial scenes, which is an important limitation in RS applications. To fill this gap, we propose a new graph relation network (GRN) for multilabel RS scene categorization. Our GRN is able to model the relations between samples (or scenes) by making use of a graph structure which is fed into network learning. For this purpose, we define a new loss function called scalable neighbor discriminative loss with binary cross entropy (SNDL-BCE) that is able to embed the graph structures through the networks more effectively. The proposed approach can guide deep learning techniques (such as convolutional neural networks) to a more discriminative metric space, where semantically similar RS scenes are closely embedded and dissimilar images are separated from a novel multilabel viewpoint. To achieve this goal, our GRN jointly maximizes a weighted leave-one-out K-nearest neighbors (K NN) score in the training set, where the weight matrix describes the contributions of the nearest neighbors associated with each RS image on its class decision, and the likelihood of the class discrimination in

show abstract

Multilabel Conditional Random Field Classification for UAV Images

Cited by 26 publications

References 14 publications

Recurrently exploring class-wise attention in a hybrid convolutional and bidirectional LSTM network for multi-label aerial image classification

Recurrently exploring class-wise attention in a hybrid convolutional and bidirectional LSTM network for multi-label aerial image classification

A survey on video classification using action recognition

Graph Relation Network: Modeling Relations Between Scenes for Multilabel Remote-Sensing Image Classification and Retrieval

Contact Info

Product

Resources

About