检索结果-内蒙古大学图书馆

27th International Conference on Pattern Recognition, ICPR 2024

作者： Kumar Sinha, Abhishek Mishra, Deepak Moorthi, S. Manthira Signal and Image Processing Area Space Applications Center Ahmedabad India Indian Institute of Space Science and Technology Thiruvananthapuram India

ISBN: (纸本)9783031781827

Transductive zero-shot learning (TZSL) aims to minimize the domain shift between the learned and true distribution of the unseen classes by allowing access to the unpaired samples from unseen classes. While many distribution alignment based methods attempt to align both visual and semantic spaces to train the classifier, their performance is still limited by confirmation bias. Additionally, bidirectional alignment approaches are based on the strong assumption that the intrinsic dimensions of visual and semantic spaces are the same, which is rarely true. In this work, we first highlight the limitations of bidirectional alignment in terms of intrinsic dimensionality. We then present a pseudo-bidirectional approach that, without any underlying assumptions on these spaces, utilizes the learned visual-to-attribute mapping to minimize the distribution shift between learned and true unseen visual feature distributions. We further utilize an entangled loss between semantic and visual space to minimize the confirmation or uncertainty bias and improve the adversarial robustness. We, theoretically and empirically, show the performance gain in addition to the adversarial robustness under the proposed setting. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Adversarial machine learning

Deep Learning Based Novel Cascaded Approach for Skin Lesion Analysis 7th

学校读者我要写书评

暂无评论

Deep Learning Based Novel Cascaded Approach for Skin Lesion ...

7th International Conference on Computer Vision and image processing, CVIP 2022

作者： Innani, Shubham Dutande, Prasad Baheti, Bhakti Baid, Ujjwal Talbar, Sanjay Center of Excellence Signal and Image Processing Shri Guru Gobind Singhji Institute of Engineering and Technology Nanded India

ISBN: (纸本)9783031314063

Patients diagnosed with skin cancer like melanoma are prone to a high mortality rate. Automatic lesion analysis is critical in skin can¬cer diagnosis and ensures effective treatment. The computer-aided diag¬nosis of such skin cancer in dermoscopic images can significantly reduce the clinicians’ workload and help improve diagnostic accuracy. Although researchers are working extensively to address this problem, early detec¬tion and accurate identification of skin lesions remain challenging. This research focuses on a two-step framework for skin lesion segmentation followed by classification for lesion analysis. We explored the effective¬ness of deep convolutional neural network (CNN) based architectures by designing an encoder-decoder architecture for skin lesion segmenta¬tion and CNN based classification network. The proposed approaches are evaluated quantitatively in terms of the Accuracy, mean Intersection over Union(mIoU) and Dice Similarity Coefficient. Our cascaded end-to-end deep learning-based approach is the first of its kind, where the classification accuracy of the lesion is significantly improved because of prior segmentation. The code is available at https://***/shubhaminnani/skin/lesion. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： image enhancement

Total variation regularized tensor ring decomposition for OCT image denoising and super-resolution

学校读者我要写书评

暂无评论

Computers in Biology and Medicine 2024年 177卷 108591-108591页

作者： Ghaderi Daneshmand, Parisa Rabbani, Hossein Medical Image & Signal Processing Research Center School of Advanced Technologies in Medicine Isfahan University of Medical Sciences Isfahan*** Iran

This paper suggests a novel hybrid tensor-ring (TR) decomposition and first-order tensor-based total variation (FOTTV) model, known as the TRFOTTV model, for super-resolution and noise suppression of optical coherence tomography (OCT) images. OCT imaging faces two fundamental problems undermining correct OCT-based diagnosis: significant noise levels and low sampling rates to speed up the capturing process. Inspired by the effectiveness of TR decomposition in analyzing complicated data structures, we suggest the TRFOTTV model for noise suppression and super-resolution of OCT images. Initially, we extract the nonlocal 3D patches from OCT data and group them to create a third-order low-rank tensor. Subsequently, using TR decomposition, we extract the correlations among all modes of the grouped OCT tensor. Finally, FOTTV is integrated into the TR model to enhance spatial smoothness in OCT images and conserve layer structures more effectively. The proximal alternating minimization and alternative direction method of multipliers are applied to solve the obtained optimization problem. The effectiveness of the suggested method is verified by four OCT datasets, demonstrating superior visual and numerical outcomes compared to state-of-the-art procedures. © 2024 Elsevier Ltd

关键词： Optical tomography

On Visualizing Music Storage Media for Modern Access to Historic Sources

学校读者我要写书评

暂无评论

On Visualizing Music Storage Media for Modern Access to Hist...

24th Eurographics Conference on Visualization, EuroVis 2022

作者： Khulusi, R. Fricke, H. Image and Signal Processing Group Leipzig University Germany Research Center Digital Organology Leipzig University Germany

ISBN: (纸本)9783038681854

Finding a balance between conserving historic objects and using them for research is one of the big issues in historic collections. Digitization holds the opportunity to offer a safe and non-destructible access to historic objects, making them available for research. With this poster, we want to give insight into our planned visualization system, using close and distant reading access for visual analysis approaches and allowing musicologists novel approaches to normally fragile and endangered media. © 2022 The Author(s) Eurographics Proceedings © 2022 The Eurographics Association.

关键词：

Geometric Correction and Mosaic Generation of Geo High Resolution Camera images

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Garg, Ankur Thapa, Nitesh Sangar, Ghansham Gaur, Neha Sarkar, Meenakshi Moorthi, S. Manthira Dhar, Debajyoti Signal and Image Processing Group Space Applications Center Indian Space Research Organisation Ahmedabad380015 India

The Geo High Resolution Camera (GHRC) aboard the GSAT-29 satellite, operated by the Indian Space Research Organization (ISRO), is a cutting-edge 6-band Visible and Near Infrared (VNIR) imager positioned in geostationary orbit at 55°E longitude. It achieves a ground sampling distance of 55 meters at nadir and covers an instantaneous area of 110x110 km, capturing imagery of the entire Earth’s disk thanks to its scan mirror mechanism. To comprehensively cover India, the GHRC employs a two-dimensional raster scanning technique, resulting in over 1,000 scenes that need to be skillfully stitched together into a seamless mosaic. This paper outlines the geolocation model and identifies potential sources of error affecting targeting accuracy. It also conducts an evaluation of the location accuracies. Challenges related to inter-band registration and inter-frame mosaicing with the GHRC are addressed. Several algorithms have been devised to address geometric correction, band-to-band registration, and seamless mosaic generation. Notably, adjustments to the instrument’s interior alignment angles, guided by ground reference images during in-flight geometric calibration, have significantly improved pointing and location precision. A groundbreaking backtracking algorithm has been developed to rectify frame-to-frame mosaicing errors for large-scale mosaic generation. This algorithm harnesses a combination of geometric models, advanced image processing techniques, and space resection. As a result of these advancements, operational generation of full India mosaics with 100-meter resolution and high geometric fidelity is now a reality. These accomplishments mark substantial progress in the capabilities of the GHRC and hold profound implications for applications in Earth observation and monitoring. Copyright © 2024, The Authors. All rights reserved.

关键词： Mirrors

Advancements in Data processing and Calibration for the Hyperspectral Imaging Satellite (HySIS)

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Garg, Ankur Patil, Abhishek Sarkar, Meenakshi Moorthi, S. Manthira Dhar, Debajyoti Signal and Image Processing Group Space Applications Center Indian Space Research Organisation Ahmedabad380015 India

Hyperspectral imaging stands as a powerful tool in Earth exploration, owing to its exceptional ability to discern intricate spectral features. India has harnessed this capability through the launch of a dedicated hyperspectral earth observation satellite. This satellite is designed to scrutinize the Earth’s surface across a broad electromagnetic spectrum range spanning from 400 nm to 2500 nm, capturing an extensive array of spectral information. One of the key strengths of this satellite lies in its impressive spatial sampling resolution of 30 meters. This high resolution further amplifies its imaging prowess, enabling it to capture detailed and accurate data about the Earth’s surface features. This paper offers a comprehensive account of the meticulous calibration procedures that were meticulously executed during the pre-launch phase. Calibration is a critical process in ensuring the accuracy and reliability of the data captured by the satellite. This involved a series of precise adjustments and measurements to fine-tune the satellite’s imaging instruments, ensuring that they were operating at their optimal levels. The paper also highlights the achieved performance metrics, showcasing the satellite’s capability to deliver high-quality, detailed imagery of the Earth’s surface. These metrics serve as a testament to the effectiveness of the calibration procedures carried out prior to the satellite’s launch. Furthermore, the paper outlines the rigorous in-orbit radiometric and geometric calibration and validation endeavors that were conducted during the commissioning phase of the mission. This phase involved further fine-tuning and validation of the satellite’s instruments in the actual space environment. It ensured that the satellite was operating accurately and reliably in orbit, ready to fulfill its mission objectives. In addition, the paper sheds light on the observations made and anomalies encountered during the in-orbit phase. These observations provide valuable

关键词： Network security

Hyperspectral Spatial Super-Resolution using Keystone Error

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Garg, Ankur Sarkar, Meenakshi Moorthi, S. Manthira Dhar, Debajyoti Signal and Image Processing Group Space Applications Center Indian Space Research Organisation Ahmedabad380015 India

Hyperspectral images aid in precise identification of objects on ground by measuring their signatures with very fine spectral resolution. If such images have high spatial resolution, it further boosts this ability. Increasing the spatial resolution with hardware by making use of bigger telescopes is a very costly affair in terms of the resources needed which makes the overall system optimization suffer. In comparison, increasing the spatial resolution with ground processing is the most optimal solution to handle this obstacle. Generally, to get advantage of both high spectral and spatial resolution data, hypersharpening approaches are used to merge high spectral resolution data with high spatial resolution. The approach gives good results when the two datasets have been acquired within very short time frame leading to similar atmospheric, viewing imaging conditions. This is usually possible when both high spectral and spatial resolution sensors are imaging from same platform. But using datasets from different dates and time often leads to radiometric artifacts which degrade the final image quality. In this work, we propose an approach to increase the spatial resolution of hyperspectral data without making use of any auxiliary input. The two key inputs for increasing the spatial resolution via superresolution algorithms are the high resolution point spread function and the sub-pixel shifts. For grating based hyperspectral sensors, misregistration between spectral channels occur due to instrument intrinsic keystone error which leads to sub-pixel shifts between spectral channels that vary across the spatial field of view. But the high resolution blur for such systems is seldom known due to the dynamic nature of the intervening atmosphere and platform jitter. Also, the optics blur for each channel differs due to differing wavelengths. To handle this dynamic effect, we first estimate the point spread function from lower resolution image using blind deconvolution and corr

关键词： Spectral resolution

EVERYONE CAN ATTACK: REPURPOSE LOSSY COMPRESSION AS A NATURAL BACKDOOR ATTACK

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yang, Sze Jue Nguyen, Quang H. Chan, Chee Seng Doan, Khoa D. College of Engineering and Computer Science VinUniversity Viet Nam Center of Image and Signal Processing Universiti Malaya Malaysia

The vulnerabilities to backdoor attacks have recently threatened the trustworthiness of machine learning models in practical applications. Conventional wisdom suggests that not everyone can be an attacker since the process of designing the trigger generation algorithm often involves significant effort and extensive experimentation to ensure the attack’s stealthiness and effectiveness. Alternatively, this paper shows that there exists a more severe backdoor threat: anyone can exploit an easily-accessible algorithm for silent backdoor attacks. Specifically, this attacker can employ the widely-used lossy image compression from a plethora of compression tools to effortlessly inject a trigger pattern into an image without leaving any noticeable trace;i.e., the generated triggers are natural artifacts. One does not require extensive knowledge to click on the "convert" or "save as" button while using tools for lossy image compression. Via this attack, the adversary does not need to design a trigger generator as seen in prior works and only requires poisoning the data. Empirically, the proposed attack consistently achieves 100% attack success rate in several benchmark datasets such as MNIST, CIFAR-10, GTSRB and CelebA. More significantly, the proposed attack can still achieve almost 100% attack success rate with very small (approximately 10%) poisoning rates in the clean label setting. The generated trigger of the proposed attack using one lossy compression algorithm is also transferable across other related compression algorithms, exacerbating the severity of this backdoor threat. This work takes another crucial step toward understanding the extensive risks of backdoor attacks in practice, urging practitioners to investigate similar attacks and relevant backdoor mitigation methods. © 2023, CC BY-NC-ND.

关键词： image compression

Modified ResNet for Volumetric Segmentation of Liver and Its Tumor from 3D CT

学校读者我要写书评

暂无评论

Modified ResNet for Volumetric Segmentation of Liver and Its...

International Conference on Artificial Intelligence and Sustainable Engineering, AISE 2020

作者： Mourya, Gajendra Kumar Bhatia, Dinesh Gogoi, Manashjit Talbar, S.N. Baid, Ujjwal Dudante, Prasad Department of Biomedical Engineering North-Eastern Hill University Meghalaya Shillong India Center of Excellence in Signal and Image Processing Shri Guru Gobind Singhji Institute of Engineering and Technology Maharashtra Nanded India

ISBN: (纸本)9789811685415

Liver tumor segmentation is the first step to investigate severity of liver disorder. The conventional automatic segmentation methods highly depend on hand-crafted features and priori information, which are tedious and time consuming for medical images. Fully convolutional network (FCN) of the deep learning model can extract features automatically and hierarchically learning concerning layers depth. In this article, we demonstrated the application of modified ResNet for the automatic segmentation of liver and tumor segmentation from 3D CT medical imaging datasets. The FCN ResNet architecture is modified while changing the levels at encoder and decoder ends to overcome the limitation of the conventional segmentation technique. The network input and output image dimension kept the same for training and testing for the designed model. LiTS and 3Dirac data use for training and testing the model. Multifold training datasets strategy utilized. Results were evaluated quantitatively and achieved DICE 0.96 and 0.81 for liver and tumor, respectively. The modified technique will assist to interventional radiologist workflow and faster the diagnostic and therapeutic procedures. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Computerized tomography

Decentralized Kernel Ridge Regression Based on Data-Dependent Random Feature

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Yang, Ruikai He, Fan He, Mingzhen Yang, Jie Huang, Xiaolin The Institute of Image Processing and Pattern Recognition The MOE Key Laboratory of System Control and Information Processing Shanghai Jiao Tong University Shanghai200240 China The STADIUS Center for Dynamical Systems Signal Processing and Data Analytics KU Leuven LeuvenB-3001 Belgium

Random feature (RF) has been widely used for node consistency in decentralized kernel ridge regression (KRR). Currently, the consistency is guaranteed by imposing constraints on coefficients of features, necessitating that the random features on different nodes are identical. However, in many applications, data on different nodes varies significantly on the number or distribution, which calls for adaptive and data-dependent methods that generate different RFs. To tackle the essential difficulty, we propose a new decentralized KRR algorithm that pursues consensus on decision functions, which allows great flexibility and well adapts data on nodes. The convergence is rigorously given and the effectiveness is numerically verified: by capturing the characteristics of the data on each node, while maintaining the same communication costs as other methods, we achieved an average regression accuracy improvement of 25.5% across six real-world data sets. Copyright © 2024, The Authors. All rights reserved.

关键词： Consensus algorithm