Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104309
Preet Chandan Kaur , Dr. Leena Ragha
Video summarization plays an important role in multiple applications by compressing lengthy video content into compressed representation. The purpose is to present a fine-tuned deep model for lecture audio video summarization. Initially, the input lecture audio-visual video is taken from the dataset. Then, the video shot segmentation (slide segmentation) is done using the YCbCr space colour model. From each video shot, the audio and video within the video shot are segmented using the Honey Badger-based Bald Eagle Algorithm (HBBEA). The HBBEA is obtained by combining the Bald Eagle Search (BES) and Honey Badger Algorithm (HBA). The DRN training is executed by HBBEA to select the finest DRN weights. The relevant video frames are merged with the audio. The proposed HBBEA-based DRN outperformed with a better F1-Score of 91.9 %, Negative predictive value (NPV) of 89.6 %, Positive predictive value (PPV) of 90.7 %, Accuracy of 91.8 %, precision of 91 %, and recall of 92.8 %.
{"title":"Optimized deep learning enabled lecture audio video summarization","authors":"Preet Chandan Kaur , Dr. Leena Ragha","doi":"10.1016/j.jvcir.2024.104309","DOIUrl":"10.1016/j.jvcir.2024.104309","url":null,"abstract":"<div><div>Video summarization plays an important role in multiple applications by compressing lengthy video content into compressed representation. The purpose is to present a fine-tuned deep model for lecture audio video summarization. Initially, the input lecture audio-visual video is taken from the dataset. Then, the video shot segmentation (slide segmentation) is done using the YCbCr space colour model. From each video shot, the audio and video within the video shot are segmented using the Honey Badger-based Bald Eagle Algorithm (HBBEA). The HBBEA is obtained by combining the Bald Eagle Search (BES) and Honey Badger Algorithm (HBA). The DRN training is executed by HBBEA to select the finest DRN weights. The relevant video frames are merged with the audio. The proposed HBBEA-based DRN outperformed with a better F1-Score of 91.9 %, Negative predictive value (NPV) of 89.6 %, Positive predictive value (PPV) of 90.7 %, Accuracy of 91.8 %, precision of 91 %, and recall of 92.8 %.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104309"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142534822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104317
Zhouyan He , Renzhi Hu , Jun Wu , Ting Luo , Haiyong Xu
For the existing encoder-noise-decoder (END) based watermarking models, since the coupling between the encoder and the decoder is weak, the encoder generally embeds certain redundant features into the cover image to enable the decoder to extract watermark completely, which will affect watermarking invisibility. To address this problem, this paper proposes a Transformer-based invertible neural network (INN) for robust image watermarking (IWFormer). In order to effectively reduce redundant features, the INN framework is utilized for the watermark embedding and extracting processes, so that the encoded features are highly consistent with the features required for decoding. For enhancing watermarking robustness, an affine Transformer module is designed by mining the global correlation of the cover image. In addition, considering that the human visual system is sensitive to low-frequency variations, the wavelet low-frequency sub-band loss is deployed to guide watermark to be embedded in middle- and high-frequency components, thus further increasing the quality of the encoded images. Experimental results demonstrate that compared with the existing state-of-the-art watermarking models, the proposed IWFormer owns remarkable advantages in terms of both watermarking invisibility and robustness.
{"title":"A Transformer-based invertible neural network for robust image watermarking","authors":"Zhouyan He , Renzhi Hu , Jun Wu , Ting Luo , Haiyong Xu","doi":"10.1016/j.jvcir.2024.104317","DOIUrl":"10.1016/j.jvcir.2024.104317","url":null,"abstract":"<div><div>For the existing encoder-noise-decoder (END) based watermarking models, since the coupling between the encoder and the decoder is weak, the encoder generally embeds certain redundant features into the cover image to enable the decoder to extract watermark completely, which will affect watermarking invisibility. To address this problem, this paper proposes a Transformer-based invertible neural network (INN) for robust image watermarking (IWFormer). In order to effectively reduce redundant features, the INN framework is utilized for the watermark embedding and extracting processes, so that the encoded features are highly consistent with the features required for decoding. For enhancing watermarking robustness, an affine Transformer module is designed by mining the global correlation of the cover image. In addition, considering that the human visual system is sensitive to low-frequency variations, the wavelet low-frequency sub-band loss is deployed to guide watermark to be embedded in middle- and high-frequency components, thus further increasing the quality of the encoded images. Experimental results demonstrate that compared with the existing state-of-the-art watermarking models, the proposed IWFormer owns remarkable advantages in terms of both watermarking invisibility and robustness.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104317"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142446136","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To safeguard the identity and copyright of a patient’s medical documents, watermarking strategies are widely used. This work provides a new dual image-based watermarking approach using the quorum function (QF) and AD interpolation technique. AD interpolation is used to create the dual images which helps to increase the embedding capacity. Moreover, the rules for using the QF are designed in such a way, that the original bits are least affected after embedding. As a result, it increases the visual quality of the stego images. A shared secret key has been employed to protect the information hidden in the medical image and to maintain the privacy and confidentiality. The experimental result using PSNR, SSIM, NCC, and EC shows that the suggested technique gives an average PSNR of 68.44 dB and SSIM is close to 0.99 after inserting 786432 watermark bits, which demonstrates the superiority of the scheme over other state-of-the-art schemes.
{"title":"A robust watermarking approach for medical image authentication using dual image and quorum function","authors":"Ashis Dey , Partha Chowdhuri , Pabitra Pal , Utpal Nandi","doi":"10.1016/j.jvcir.2024.104299","DOIUrl":"10.1016/j.jvcir.2024.104299","url":null,"abstract":"<div><div>To safeguard the identity and copyright of a patient’s medical documents, watermarking strategies are widely used. This work provides a new dual image-based watermarking approach using the quorum function (QF) and AD interpolation technique. AD interpolation is used to create the dual images which helps to increase the embedding capacity. Moreover, the rules for using the QF are designed in such a way, that the original bits are least affected after embedding. As a result, it increases the visual quality of the stego images. A shared secret key has been employed to protect the information hidden in the medical image and to maintain the privacy and confidentiality. The experimental result using PSNR, SSIM, NCC, and EC shows that the suggested technique gives an average PSNR of 68.44 dB and SSIM is close to 0.99 after inserting 786432 watermark bits, which demonstrates the superiority of the scheme over other state-of-the-art schemes.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104299"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142356896","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104297
Xiuya Shi , Yi Yang , Hao Liu , Litai Ma , Zhibo Zhao , Chao Ren
Low-dose computed tomography (LDCT) is an emerging medical diagnostic tool that reduces radiation exposure but suffers from noise retention. Current CNN-based LDCT denoising algorithms struggle to capture comprehensive global representations, impacting diagnostic accuracy. To address this, we propose a novel Hierarchical Prior-guided Iterative Denoising Network (HPIDN) for LDCT images, consisting of two main modules: the Dynamic Feature Extraction and Fusion Module (DFEFM) and the Feature-domain Iterative Denoising Module (FIDM). DFEFM dynamically captures a comprehensive representation, encompassing detailed local features in intra-relationships and complex global features in inter-relationships. It effectively guides the multi-stage iterative denoising process. FIDM hierarchically fuses the prior with image features from DFEFM by using the dual-domain attention fusion sub-network (DAFSN), enhancing denoising robustness and adaptability. This yields higher-quality images with reduced noise artifacts. Extensive experiments on the Mayo and ELCAP Datasets demonstrate the superior performance of our method quantitatively and qualitatively, improving diagnostic accuracy of lung diseases.
{"title":"HPIDN: A Hierarchical prior-guided iterative denoising network with global–local fusion for enhancing low-dose CT images","authors":"Xiuya Shi , Yi Yang , Hao Liu , Litai Ma , Zhibo Zhao , Chao Ren","doi":"10.1016/j.jvcir.2024.104297","DOIUrl":"10.1016/j.jvcir.2024.104297","url":null,"abstract":"<div><div>Low-dose computed tomography (LDCT) is an emerging medical diagnostic tool that reduces radiation exposure but suffers from noise retention. Current CNN-based LDCT denoising algorithms struggle to capture comprehensive global representations, impacting diagnostic accuracy. To address this, we propose a novel Hierarchical Prior-guided Iterative Denoising Network (HPIDN) for LDCT images, consisting of two main modules: the Dynamic Feature Extraction and Fusion Module (DFEFM) and the Feature-domain Iterative Denoising Module (FIDM). DFEFM dynamically captures a comprehensive representation, encompassing detailed local features in intra-relationships and complex global features in inter-relationships. It effectively guides the multi-stage iterative denoising process. FIDM hierarchically fuses the prior with image features from DFEFM by using the dual-domain attention fusion sub-network (DAFSN), enhancing denoising robustness and adaptability. This yields higher-quality images with reduced noise artifacts. Extensive experiments on the Mayo and ELCAP Datasets demonstrate the superior performance of our method quantitatively and qualitatively, improving diagnostic accuracy of lung diseases.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104297"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142424087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104306
Yaqi Zhao, Yue Li
Medical ultrasound imaging is a widely used non-invasive method for diagnosing diseases. However, these images contain significant speckle noise, which differs from the characteristics of natural images. This makes effective lossless compression of medical ultrasound images a challenging task. In this paper, we propose a novel hybrid ultrasound image lossless learning compression framework. Firstly, we use the traditional DCT (discrete cosine transform) to transform the original raw pixels of ultrasound images into the frequency domain. Secondly, to effectively compress the numerical values in the frequency domain, we decompose the DCT coefficients into different groups to reduce local and global information redundancy in the frequency domain. Finally, we use learned and non-learned methods to compress the DCT coefficients of different groups separately. The experimental results show that on the Breast ultrasound image dataset, our proposed method achieves a bit rate reduction of 8.6% to 68.9% compared to learned and non-learned methods.
{"title":"Lossless medical ultrasound image compression based on frequency domain decomposition","authors":"Yaqi Zhao, Yue Li","doi":"10.1016/j.jvcir.2024.104306","DOIUrl":"10.1016/j.jvcir.2024.104306","url":null,"abstract":"<div><div>Medical ultrasound imaging is a widely used non-invasive method for diagnosing diseases. However, these images contain significant speckle noise, which differs from the characteristics of natural images. This makes effective lossless compression of medical ultrasound images a challenging task. In this paper, we propose a novel hybrid ultrasound image lossless learning compression framework. Firstly, we use the traditional DCT (discrete cosine transform) to transform the original raw pixels of ultrasound images into the frequency domain. Secondly, to effectively compress the numerical values in the frequency domain, we decompose the DCT coefficients into different groups to reduce local and global information redundancy in the frequency domain. Finally, we use learned and non-learned methods to compress the DCT coefficients of different groups separately. The experimental results show that on the Breast ultrasound image dataset, our proposed method achieves a bit rate reduction of 8.6% to 68.9% compared to learned and non-learned methods.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104306"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142424088","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104312
Seunggyun Woo , Keunsoo Ko , Chang-Su Kim
In general, CNN-based inpainting can recover local patterns effectively using convolutional filters, but it may not exploit global correlation fully. On the other hand, transformer-based inpainting can fill in large holes faithfully based on global correlation, rather than local one. In this paper, we propose a novel image inpainting algorithm, called local and global mixture (LGM), to take advantage of the strengths of both approaches and compensate for their weaknesses. The LGM network comprises the local inpainting network (LIN) and the global inpainting network (GIN) in parallel, which are based on convolutional layers and transformer blocks, respectively, and exchange their intermediate results with each other. Furthermore, we develop an error propagation model with a continuous error mask, updated in LIN but used in both LIN and GIN to provide more reliable inpainting results. Extensive experiments demonstrate that the proposed LGM algorithm provides excellent inpainting performance, which indicates the efficacy of the parallel combination of LIN and GIN and the effectiveness of the error propagation model.
一般来说，基于 CNN 的涂色可以利用卷积滤波器有效地恢复局部模式，但可能无法充分利用全局相关性。另一方面，基于变换器的内绘可以基于全局相关性而非局部相关性忠实地填补大漏洞。在本文中，我们提出了一种名为局部和全局混合（LGM）的新型图像内绘算法，以利用这两种方法的优势并弥补它们的不足。LGM 网络由本地 Inpainting 网络 (LIN) 和全局 Inpainting 网络 (GIN) 并行组成，这两个网络分别基于卷积层和变换块，并相互交换中间结果。此外，我们还开发了一种带有连续误差掩码的误差传播模型，该模型在 LIN 中更新，但同时用于 LIN 和 GIN，以提供更可靠的绘制结果。广泛的实验证明，所提出的 LGM 算法具有出色的内绘制性能，这表明了 LIN 和 GIN 并行组合的功效以及误差传播模型的有效性。
{"title":"Local and global mixture network for image inpainting","authors":"Seunggyun Woo , Keunsoo Ko , Chang-Su Kim","doi":"10.1016/j.jvcir.2024.104312","DOIUrl":"10.1016/j.jvcir.2024.104312","url":null,"abstract":"<div><div>In general, CNN-based inpainting can recover local patterns effectively using convolutional filters, but it may not exploit global correlation fully. On the other hand, transformer-based inpainting can fill in large holes faithfully based on global correlation, rather than local one. In this paper, we propose a novel image inpainting algorithm, called local and global mixture (LGM), to take advantage of the strengths of both approaches and compensate for their weaknesses. The LGM network comprises the local inpainting network (LIN) and the global inpainting network (GIN) in parallel, which are based on convolutional layers and transformer blocks, respectively, and exchange their intermediate results with each other. Furthermore, we develop an error propagation model with a continuous error mask, updated in LIN but used in both LIN and GIN to provide more reliable inpainting results. Extensive experiments demonstrate that the proposed LGM algorithm provides excellent inpainting performance, which indicates the efficacy of the parallel combination of LIN and GIN and the effectiveness of the error propagation model.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104312"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142534823","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104313
Jie Yang, Yuantong Zhang, Zhenzhong Chen, Daiqin Yang
Exposure problems, including underexposure and overexposure, can significantly degrade image quality. Poorly exposed images often suffer from coupled illumination degradation and detail degradation, aggravating the difficulty of recovery. These necessitate a spatial discriminating exposure correction, making achieving uniformly exposed and visually consistent images challenging. To address these issues, we propose an Illumination-guided Dual-domain Network (IDNet), which employs a Dual-Domain Module (DDM) to simultaneously recover illumination and details from the frequency and spatial domains, respectively. The DDM also integrates a structural re-parameterization technique to enhance the detail-aware capabilities with reduced computational cost. An Illumination Mask Predictor (IMP) is introduced to guide exposure correction by estimating the optimal illumination mask. The comparison with 26 methods on three benchmark datasets shows that IDNet achieves superior performance with fewer parameters and lower computational complexity. These results confirm the effectiveness and efficiency of our approach in enhancing image quality across various exposure scenarios.
{"title":"An illumination-guided dual-domain network for image exposure correction","authors":"Jie Yang, Yuantong Zhang, Zhenzhong Chen, Daiqin Yang","doi":"10.1016/j.jvcir.2024.104313","DOIUrl":"10.1016/j.jvcir.2024.104313","url":null,"abstract":"<div><div>Exposure problems, including underexposure and overexposure, can significantly degrade image quality. Poorly exposed images often suffer from coupled illumination degradation and detail degradation, aggravating the difficulty of recovery. These necessitate a spatial discriminating exposure correction, making achieving uniformly exposed and visually consistent images challenging. To address these issues, we propose an Illumination-guided Dual-domain Network (IDNet), which employs a Dual-Domain Module (DDM) to simultaneously recover illumination and details from the frequency and spatial domains, respectively. The DDM also integrates a structural re-parameterization technique to enhance the detail-aware capabilities with reduced computational cost. An Illumination Mask Predictor (IMP) is introduced to guide exposure correction by estimating the optimal illumination mask. The comparison with 26 methods on three benchmark datasets shows that IDNet achieves superior performance with fewer parameters and lower computational complexity. These results confirm the effectiveness and efficiency of our approach in enhancing image quality across various exposure scenarios.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104313"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142446137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104300
Xinyi Huang, Hongxia Wang
The wide spread of digital documents makes it essential to protect intellectual property and information security. As a key method of digital copyright protection, robust document watermarking technology has attracted much attention in this context. With the rapid development of current electronic devices, the ways of document theft are no longer limited to copy and transmission. Due to the convenient and fast shooting operation of the camera on paper or screen, current text watermarking methods need to be robust to cope with cross-media transmission. To realize the corresponding robust text watermarking, a text watermarking scheme based on the average skeleton mass of characters is proposed in this paper, and the average skeleton mass of adjacent characters is used to represent the watermark information. In this paper, a watermarking scheme is designed to modify character pixels, which can modify glyphs without loss of transparency and provide high embedding capacity. Compared with the existing manually designed font-based text watermarking schemes, this scheme does not need to accurately segment characters, nor does it rely on stretching characters to the same size for matching, which reduces the need for character segmentation. In addition, the experimental results show that the proposed watermarking scheme can be robust to the information transmission modes including print-scan, print-camera and screen-camera.
{"title":"Robust text watermarking based on average skeleton mass of characters against cross-media attacks","authors":"Xinyi Huang, Hongxia Wang","doi":"10.1016/j.jvcir.2024.104300","DOIUrl":"10.1016/j.jvcir.2024.104300","url":null,"abstract":"<div><div>The wide spread of digital documents makes it essential to protect intellectual property and information security. As a key method of digital copyright protection, robust document watermarking technology has attracted much attention in this context. With the rapid development of current electronic devices, the ways of document theft are no longer limited to copy and transmission. Due to the convenient and fast shooting operation of the camera on paper or screen, current text watermarking methods need to be robust to cope with cross-media transmission. To realize the corresponding robust text watermarking, a text watermarking scheme based on the average skeleton mass of characters is proposed in this paper, and the average skeleton mass of adjacent characters is used to represent the watermark information. In this paper, a watermarking scheme is designed to modify character pixels, which can modify glyphs without loss of transparency and provide high embedding capacity. Compared with the existing manually designed font-based text watermarking schemes, this scheme does not need to accurately segment characters, nor does it rely on stretching characters to the same size for matching, which reduces the need for character segmentation. In addition, the experimental results show that the proposed watermarking scheme can be robust to the information transmission modes including print-scan, print-camera and screen-camera.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104300"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142356790","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104296
Bindu Puthentharayil Vikraman , Jabeena Afthab
Nowadays, image compression is gaining popularity in various fields because of its storage and transmission capability. This work aims to introduce a medical image (MI) compression model in brain magnetic resonance images (MRI) to mitigate issues in bandwidth and storage. Initially, pre-processing is done to neglect the noises in inputs using the Adaptive Linear Smoothing and Histogram Equalization (ALSHE) method. Then, the Region of Interest (ROI) and Non-ROI parts are separately segmented by the Optimized Fuzzy C-Means (OFCM) approach for reducing high complexity issues. Finally, a novel Hybrid Discrete Cosine Transform-Improved Zero Wavelet (DCT-IZW) is proposed for lossless compression and Hybrid Equilibrium Optimization-Capsule Auto Encoder (EO-CAE) for lossy compression. Then, the compressed ROI and Non-ROI images are added together, and the inverse operation of the compression process is performed to obtain the reconstructed image. This study used BRATS (2015, 2018) datasets for simulation and attained better performance than other existing methods.
如今，图像压缩因其存储和传输能力强而在各个领域越来越受欢迎。这项工作旨在引入脑磁共振图像（MRI）中的医学图像压缩模型，以缓解带宽和存储问题。首先，使用自适应线性平滑和直方图均衡（ALSHE）方法进行预处理，以忽略输入中的噪声。然后，使用优化模糊 C-Means (OFCM) 方法分别分割感兴趣区域 (ROI) 和非感兴趣区域 (ROI) 部分，以减少高复杂性问题。最后，提出了用于无损压缩的新型混合离散余弦变换-改进零小波（DCT-IZW）和用于有损压缩的混合平衡优化-胶囊自动编码器（EO-CAE）。然后，将压缩后的 ROI 和非 ROI 图像相加，并对压缩过程进行逆运算，得到重建图像。该研究使用 BRATS（2015、2018）数据集进行模拟，取得了比其他现有方法更好的性能。
{"title":"Effective image compression using hybrid DCT and hybrid capsule auto encoder for brain MR images","authors":"Bindu Puthentharayil Vikraman , Jabeena Afthab","doi":"10.1016/j.jvcir.2024.104296","DOIUrl":"10.1016/j.jvcir.2024.104296","url":null,"abstract":"<div><div>Nowadays, image compression is gaining popularity in various fields because of its storage and transmission capability. This work aims to introduce a medical image (MI) compression model in brain magnetic resonance images (MRI) to mitigate issues in bandwidth and storage. Initially, pre-processing is done to neglect the noises in inputs using the Adaptive Linear Smoothing and Histogram Equalization (ALSHE) method. Then, the Region of Interest (ROI) and Non-ROI parts are separately segmented by the Optimized Fuzzy C-Means (OFCM) approach for reducing high complexity issues. Finally, a novel Hybrid Discrete Cosine Transform-Improved Zero Wavelet (DCT-IZW) is proposed for lossless compression and Hybrid Equilibrium Optimization-Capsule Auto Encoder (EO-CAE) for lossy compression. Then, the compressed ROI and Non-ROI images are added together, and the inverse operation of the compression process is performed to obtain the reconstructed image. This study used BRATS (2015, 2018) datasets for simulation and attained better performance than other existing methods.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104296"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142356791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-10-01DOI: 10.1016/j.jvcir.2024.104307
Jixin Liu, Shabo Hu, Haigen Yang, Ning Sun
For video intelligence applications in private scenes such as home environments, traditional image processing methods are usually based on clear raw data and are prone to privacy leakage. Therefore, our team proposed multilayer compressed sensing (MCS) encoding to reduce image quality for visual privacy protection (VPP). However, the way in which MCS coding is implemented leads to unavoidable information loss. On this basis, inspired by the image quilting (IQ) algorithm, an image quilting heuristic MCS (IQ-MCS) coding method is proposed in this paper to improve the problem of faster information loss in the MCS coding process, which means that a similar privacy protection effect is achieved at lower coding layers, thus obtaining better application performance. To evaluate the level of VPP, a VPP evaluation algorithm is proposed that is more in line with subjective assessment. Finally, a correlation model between the VPP level and the performance of smart applications is established to balance the relationships between them, taking the detection of abnormal human behavior in private scenes as an example. The model can also provide a reference for the evaluation of other privacy protection methods.
{"title":"Image quilting heuristic compressed sensing video privacy protection coding for abnormal behavior detection in private scenes","authors":"Jixin Liu, Shabo Hu, Haigen Yang, Ning Sun","doi":"10.1016/j.jvcir.2024.104307","DOIUrl":"10.1016/j.jvcir.2024.104307","url":null,"abstract":"<div><div>For video intelligence applications in private scenes such as home environments, traditional image processing methods are usually based on clear raw data and are prone to privacy leakage. Therefore, our team proposed multilayer compressed sensing (MCS) encoding to reduce image quality for visual privacy protection (VPP). However, the way in which MCS coding is implemented leads to unavoidable information loss. On this basis, inspired by the image quilting (IQ) algorithm, an image quilting heuristic MCS (IQ-MCS) coding method is proposed in this paper to improve the problem of faster information loss in the MCS coding process, which means that a similar privacy protection effect is achieved at lower coding layers, thus obtaining better application performance. To evaluate the level of VPP, a VPP evaluation algorithm is proposed that is more in line with subjective assessment. Finally, a correlation model between the VPP level and the performance of smart applications is established to balance the relationships between them, taking the detection of abnormal human behavior in private scenes as an example. The model can also provide a reference for the evaluation of other privacy protection methods.</div></div>","PeriodicalId":54755,"journal":{"name":"Journal of Visual Communication and Image Representation","volume":"104 ","pages":"Article 104307"},"PeriodicalIF":2.6,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142424084","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号