Fire Image Generation Based on ACGAN

Zhikai, Yang; Bu, Leping; Wang, Teng; Tianrui, Zheng; Fen, WU

doi:10.1109/ccdc.2019.8832678

Cited by 6 publications

(4 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The use of GANs for fire imagery generation is a nascent field, with works such as the ones proposed by Zhikai et al [43], Park et al [26], and Zhikai et al [41] focusing on the generation of synthetic visible fire images to augment datasets used mainly for the training of fire detection or segmentation methods.…”

Section: Gans For Synthesizing Ir Imagery For Image Analysismentioning

confidence: 99%

“…For the domain of synthetic fire image generation, there has been recent work regarding the use of GANs for the task of visible fire image synthesis [41,26,43], wildfire smoke generation [18], and artificial wildfire infrared (IR) image generation and fusion [6]. In particular, the work by Ciprián-Sánchez et al [6] is of special relevance to the present paper as it is, to the best of our knowledge, the first one to employ GANs to generate artificial IR fire images from visible ones.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Computer Vision-based Characterization of Large-scale Jet Flames using a Synthetic Infrared Image Generation Approach

Pérez-Guerrero¹,

Ciprián-Sánchez²,

Palacios³

et al. 2022

Preprint

View full text Add to dashboard Cite

Among the different kinds of fire accidents that can occur during industrial activities that involve hazardous materials, jet fires are one of the lesser-known types. This is because they are often involved in a process that generates a sequence of other accidents of greater magnitude, known as domino effect. Flame impingement usually causes domino effects, and jet fires present specific features that can significantly increase the probability of this happening. These features become relevant from a risk analysis perspective, making their proper characterization a crucial task. Deep Learning approaches have become extensively used for tasks such as jet fire characterization; however, these methods are heavily dependent on the amount of data and the quality of the labels. Data acquisition of jet fires involve expensive experiments, especially so if infrared imagery is used. Therefore, this paper proposes the use of Generative Adversarial Networks to produce plausible infrared images from visible ones, making experiments less expensive and allowing for other potential applications. The results suggest that it is possible to realistically replicate the results for experiments carried out using both visible and infrared cameras. The obtained results are compared with some previous experiments, and it is shown that similar results were obtained.

show abstract

Section: Gans For Synthesizing Ir Imagery For Image Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Computer Vision-based Characterization of Large-scale Jet Flames using a Synthetic Infrared Image Generation Approach

Pérez-Guerrero¹,

Ciprián-Sánchez²,

Palacios³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Mountain-image data are easy to obtain, owing to the availability of various built-up datasets. However, not only is there a dearth of fire or smoke images of wildfires in datasets, but such data are also relatively difficult to obtain because they require the use of installed surveillance cameras or operational drones at the site of the wildfire [28,29]. Therefore, research on damage detection is frequently faced with a data imbalance problem, which causes overfitting; overfitting results in the deterioration of the model performance [30].…”

Section: Introductionmentioning

confidence: 99%

Wildfire-Detection Method Using DenseNet and CycleGAN Data Augmentation-Based Remote Camera Imagery

et al. 2020

View full text Add to dashboard Cite

To minimize the damage caused by wildfires, a deep learning-based wildfire-detection technology that extracts features and patterns from surveillance camera images was developed. However, many studies related to wildfire-image classification based on deep learning have highlighted the problem of data imbalance between wildfire-image data and forest-image data. This data imbalance causes model performance degradation. In this study, wildfire images were generated using a cycle-consistent generative adversarial network (CycleGAN) to eliminate data imbalances. In addition, a densely-connected-convolutional-networks-based (DenseNet-based) framework was proposed and its performance was compared with pre-trained models. While training with a train set containing an image generated by a GAN in the proposed DenseNet-based model, the best performance result value was realized among the models with an accuracy of 98.27% and an F1 score of 98.16, obtained using the test dataset. Finally, this trained model was applied to high-quality drone images of wildfires. The experimental results showed that the proposed framework demonstrated high wildfire-detection accuracy.

show abstract

“…Energy-based Generative Adversarial Network (EBGAN) [ 23 ] introduced the concept and method of energy into GAN and regarded the discriminator as an energy function. ACGAN [ 24 , 25 ] added an auxiliary classifier to the output of the discriminator to improve the performance of GAN, and ACGAN also proposed using the class of each sample to update and improve the loss function, which significantly improved the performance of the network model. In the field of LE behavior recognition of short-wave radio stations, ACGAN can generate some labeled signals according to a small number of labeled signals, which achieves the purpose of data augmentation.…”

Section: Introductionmentioning

confidence: 99%

Unidimensional ACGAN Applied to Link Establishment Behaviors Recognition of a Short-Wave Radio Station

Chen

Lei

2020

Sensors

View full text Add to dashboard Cite

It is difficult to obtain many labeled Link Establishment (LE) behavior signals sent by non-cooperative short-wave radio stations. We propose a novel unidimensional Auxiliary Classifier Generative Adversarial Network (ACGAN) to get more signals and then use unidimensional DenseNet to recognize LE behaviors. Firstly, a few real samples were randomly selected from many real signals as the training set of unidimensional ACGAN. Then, the new training set was formed by combining real samples with fake samples generated by the trained ACGAN. In addition, the unidimensional convolutional auto-coder was proposed to describe the reliability of these generated samples. Finally, different LE behaviors could be recognized without the communication protocol standard by using the new training set to train unidimensional DenseNet. Experimental results revealed that unidimensional ACGAN effectively augmented the training set, thus improving the performance of recognition algorithm. When the number of original training samples was 400, 700, 1000, or 1300, the recognition accuracy of unidimensional ACGAN+DenseNet was 1.92, 6.16, 4.63, and 3.06% higher, respectively, than that of unidimensional DenseNet.

show abstract

Fire Image Generation Based on ACGAN

Cited by 6 publications

References 5 publications

Computer Vision-based Characterization of Large-scale Jet Flames using a Synthetic Infrared Image Generation Approach

Computer Vision-based Characterization of Large-scale Jet Flames using a Synthetic Infrared Image Generation Approach

Wildfire-Detection Method Using DenseNet and CycleGAN Data Augmentation-Based Remote Camera Imagery

Unidimensional ACGAN Applied to Link Establishment Behaviors Recognition of a Short-Wave Radio Station

Contact Info

Product

Resources

About