Deep Learning With Weak Supervision for Disaster Scene Description in Low-Altitude Imagery

Presa-Reyes, Maria; Tao, Yudong; Chen, Shu‐Ching; Shyu, Mei‐Ling

doi:10.1109/tgrs.2021.3129443

Cited by 5 publications

(4 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of big data, raw video is a kind of unstructured data, the content of which cannot be directly understood by the computer and reflects the relevant content [ 18 ]. The structured description of the scene mainly includes three levels, which are the description of the essence of the object, the description of the attributes of the object, and the description of the attribute relationship between the objects.…”

Section: Methodsmentioning

confidence: 99%

Scene Classification in the Environmental Art Design by Using the Lightweight Deep Learning Model under the Background of Big Data

Liu¹

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

On the basis of scene visual understanding technology, the research aims to further improve the classification efficiency and classification accuracy of art design scenes. The lightweight deep learning (DL) model based on big data is used as the main method to achieve real-time detection and recognition of multiple targets and classification of the multilabel scene. This research introduces the related foundations of the DL network and the lightweight object detection involved. The data for a multilabel scene classifier are constructed and the design of the convolutional neural network (CNN) model is described. On public datasets, the effectiveness of the lightweight object detection algorithm is verified to ensure its feasibility in the classification of actual scenes. The simulation results indicate that compared with the YOLOv3-Tiny model, the improved IRDA-YOLOv3 model reduces the number of parameters by 56.2%, the amount of computation by 46.3%, and the forward computation time of the network by 0.2 ms. It means that the IRDA-YOLOv3 network obtained after the improvement can realize the lightweight of the network. In the scene classification of complex traffic roads, the classification model of the multilabel scene can predict all kinds of semantic information of a single image and the classification accuracy for the four scenes is more than 90%. In summary, the discussed classification method based on the lightweight DL model is suitable for complex practical scenes. The constructed lightweight network improves the representational ability of the network and has certain research value for scene classification problems.

show abstract

Section: Methodsmentioning

confidence: 99%

Scene Classification in the Environmental Art Design by Using the Lightweight Deep Learning Model under the Background of Big Data

Liu¹

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…Approaches [129], Ensemble Models [123] Feature-level, Decision-level, Classifierlevel [5] Various NER datasets, Scientific papers [52] Improved performance, Effectiveness in NER [125], [158] Difficulty in handling complex entities, Limited generalization [132] Fine-grained NER, Multi-modal fusion [159] Text Classification Ensemble Algorithms [160], Deep Learning Approaches [161] Feature-level, Decision-level, Classifierlevel [162] Various Text Classification datasets, Cybersecurity [158] Improved classification accuracy, Robustness to noise [163], [164] Difficulty in handling diverse data, Limited interpretability [165] Scalable ensemble techniques, Interpretable ensembles [166] Machine Translation…”

Section: Learningmentioning

confidence: 99%

“…Approaches [169], Ensemble Models [170] Feature-level, Decision-level, Classifierlevel [171] Various Question Answering datasets [172] Improved accuracy, Robustness to noise, Realtime performance [162], [164] Difficulty in handling complex questions, Limited scalability [132] Multi-modal fusion, Temporal modeling [159] to effectively extract both acoustic and linguistic features, leading to improved accuracy in speech recognition tasks.…”

Section: Learningmentioning

confidence: 99%

Ensemble Multifeatured Deep Learning Models and Applications: A Survey

Abimannan,

El-Alfy,

Chang

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Ensemble multifeatured deep learning methodology has emerged as a powerful approach to overcome the limitations of single deep learning models in terms of generalization, robustness, and performance. This survey provides an extended review of ensemble multifeatured deep learning models, and their applications, challenges, and future directions. We explore potential applications of these models across various domains, including computer vision, medical imaging, natural language processing, and speech recognition. By combining the strengths of multiple models and features, ensemble multifeatured deep learning models have demonstrated improved performance and adaptability in diverse problem settings. We also discuss the challenges associated with these models, such as model interpretability, computational complexity, ensemble model selection, adversarial robustness, and personalized and federated learning. This survey highlights recent advancements in addressing these challenges and emphasizes the importance of continued research in tackling these issues to enable widespread adoption of ensemble multifeatured deep learning models. It provides an outlook on future research directions, focusing on the development of new algorithms, frameworks, and hardware architectures that can efficiently handle the large-scale computations required by these models. Moreover, it underlines the need for a better understanding of the trade-offs between model complexity, accuracy, and computational resources to optimize the design and deployment of ensemble multifeatured deep learning models.

show abstract

“…Compared with fully supervised object detection (FSOD) [1][2][3][4][5][6][7][8], the major advantage of weakly supervised object detection (WSOD) is that only image-level category annotations are necessary for training the WSOD model. Considering the low cost of data labeling, WSOD has been widely researched in recent years [9][10][11][12][13][14][15][16][17] and has been applied in scene classification [18,19], disaster detection [20,21], military [22,23], and other applications [24][25][26][27][28][29].…”

Section: Introductionmentioning

confidence: 99%

SAM-Induced Pseudo Fully Supervised Learning for Weakly Supervised Object Detection in Remote Sensing Images

Qian,

Lin,

Chen

et al. 2024

Remote Sensing

View full text Add to dashboard Cite

Weakly supervised object detection (WSOD) in remote sensing images (RSIs) aims to detect high-value targets by solely utilizing image-level category labels; however, two problems have not been well addressed by existing methods. Firstly, the seed instances (SIs) are mined solely relying on the category score (CS) of each proposal, which is inclined to concentrate on the most salient parts of the object; furthermore, they are unreliable because the robustness of the CS is not sufficient due to the fact that the inter-category similarity and intra-category diversity are more serious in RSIs. Secondly, the localization accuracy is limited by the proposals generated by the selective search or edge box algorithm. To address the first problem, a segment anything model (SAM)-induced seed instance-mining (SSIM) module is proposed, which mines the SIs according to the object quality score, which indicates the comprehensive characteristic of the category and the completeness of the object. To handle the second problem, a SAM-based pseudo-ground truth-mining (SPGTM) module is proposed to mine the pseudo-ground truth (PGT) instances, for which the localization is more accurate than traditional proposals by fully making use of the advantages of SAM, and the object-detection heads are trained by the PGT instances in a fully supervised manner. The ablation studies show the effectiveness of the SSIM and SPGTM modules. Comprehensive comparisons with 15 WSOD methods demonstrate the superiority of our method on two RSI datasets.

show abstract

Deep Learning With Weak Supervision for Disaster Scene Description in Low-Altitude Imagery

Cited by 5 publications

References 21 publications

Scene Classification in the Environmental Art Design by Using the Lightweight Deep Learning Model under the Background of Big Data

Scene Classification in the Environmental Art Design by Using the Lightweight Deep Learning Model under the Background of Big Data

Ensemble Multifeatured Deep Learning Models and Applications: A Survey

SAM-Induced Pseudo Fully Supervised Learning for Weakly Supervised Object Detection in Remote Sensing Images

Contact Info

Product

Resources

About