检索结果-内蒙古大学图书馆

SAM-MIL: A Spatial Contextual Aware Multiple Instance Learning Approach for Whole Slide Image Classification

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Fang, Heng Huang, Sheng Tang, Wenhao Huangfu, Luwen Liu, Bo School of Big Data & Software Engineering Chongqing University Chongqing China Fowler College of Business San Diego State University San DiegoCA United States School of Computer Science and Information Engineering Hefei University of Technology Hefei China

Multiple Instance Learning (MIL) represents the predominant framework in Whole Slide Image (WSI) classification, covering aspects such as sub-typing, diagnosis, and beyond. Current MIL models predominantly rely on instance-level features derived from pretrained models such as ResNet. These models segment each WSI into independent patches and extract features from these local patches, leading to a significant loss of global spatial context and restricting the model’s focus to merely local features. To address this issue, we propose a novel MIL framework, named SAM-MIL, that emphasizes spatial contextual awareness and explicitly incorporates spatial context by extracting comprehensive, image-level information. The Segment Anything Model (SAM) represents a pioneering visual segmentation foundational model that can capture segmentation features without the need for additional fine-tuning, rendering it an outstanding tool for extracting spatial context directly from raw WSIs. Our approach includes the design of group feature extraction based on spatial context and a SAM-Guided Group Masking strategy to mitigate class imbalance issues. We implement a dynamic mask ratio for different segmentation categories and supplement these with representative group features of categories. Moreover, SAM-MIL divides instances to generate additional pseudo-bags, thereby augmenting the training set, and introduces consistency of spatial context across pseudo-bags to further enhance the model’s performance. Experimental results on the CAMELYON-16 and TCGA Lung Cancer datasets demonstrate that our proposed SAM-MIL model outperforms existing mainstream methods in WSIs classification. Our open-source implementation code is is available at https://***/FangHeng/SAM-MIL. © 2024, CC BY-NC-SA.

关键词： Self-supervised learning

Evidential Calibrated Uncertainty-Guided Interactive Segmentation paradigm for Ultrasound Images

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Shang, Jiang Wu, Yuanmeng Han, Xiaoxiang Chen, Xi Zhang, Qi Lab Shanghai Institute for Advanced Communication and Data Science School of Communication and Information Engineering Shanghai University Shanghai200444 China The Software Engineering Institute East China Normal University Shanghai200062 China The College of Liberal Arts and Science University of Illinois Urbana-Champaign UrbanaIL United States

Accurate and robust ultrasound image segmentation is critical for computer-aided diagnostic systems. Nevertheless, the inherent challenges of ultrasound imaging, such as blurry boundaries and speckle noise, often cause traditional segmentation methods to struggle with performance. Despite recent advancements in universal image segmentation, such as the Segment Anything Model, existing interactive segmentation methods still suffer from inefficiency and lack of specialization. These methods rely heavily on extensive accurate manual or random sampling prompts for interaction, necessitating numerous prompts and iterations to reach satisfactory performance. In response to this challenge, we propose the Evidential Uncertainty-Guided Interactive Segmentation (EUGIS), an end-to-end, efficient tiered interactive segmentation paradigm based on evidential uncertainty estimation for ultrasound image segmentation. Specifically, EUGIS harnesses evidence-based uncertainty estimation, grounded in Dempster-Shafer theory and Subjective Logic, to gauge the level of uncertainty in the predictions of model for different regions. By prioritizing sampling the high-uncertainty region, our method can effectively simulate the interactive behavior of well-trained radiologists, enhancing the targeted of sampling while reducing the number of prompts and iterations required. Additionally, we propose a trainable calibration mechanism for uncertainty estimation, which can further optimize the boundary between certainty and uncertainty, thereby enhancing the confidence of uncertainty estimation. Extensive experiments on three ultrasound datasets demonstrate the competitiveness of EUGIS against the state-of-the-art non-interactive segmentation and interactive segmentation models. We believe this new paradigm will provide a novel perspective for the field of interactive segmentation and is expected to promote further development of interactive image segmentation for ultrasound image. Code and data wi

关键词： Ultrasonic imaging

Stereo Matching with Cost Volume based Sparse Disparity Propagation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Xue, Wei Peng, Xiaojiang College of Big Data and Internet Shenzhen Technology University Shenzhen China College of Computer Science and Software Engineering Shenzhen University Shenzhen China

Stereo matching is crucial for binocular stereo vision. Existing methods mainly focus on simple disparity map fusion to improve stereo matching, which require multiple dense or sparse disparity maps. In this paper, we propose a simple yet novel scheme, termed as feature disparity propagation, to improve general stereo matching based on matching cost volume and sparse matching feature points. Specifically, our scheme first calculates a reliable sparse disparity map by local feature matching, and then refines the disparity map by propagating reliable disparities to neighbouring pixels in the matching cost domain. In addition, considering the gradient and multi-scale information of local disparity regions, we present a p-Census cost measure based on the well-known AD-Census, which guarantees the robustness of cost volume even without the cost aggregation step. Extensive experiments on Middlebury stereo benchmark V3 demonstrate that our scheme achieves promising performance comparable to state-of-the-art methods. Copyright © 2022, The Authors. All rights reserved.

关键词： Stereo image processing

Vision Based Lane Assistance System

学校读者我要写书评

暂无评论

Vision Based Lane Assistance System

Emerging Techniques in Computational Intelligence (ICETCI), International Conference on

作者： Monishaa Senthilkumar Logaprakash M Himajit Aithal Saravanakumar C Shanmgam Department of Artificial intelligence and data science Sri Ramakrishna Engineering college Coimbatore India Senior AI Architect Bosch Global Software Technologies Pvt Ltd. Bengaluru India Senior Tech Lead Bosch Global Software Technologies Pvt Ltd. Coimbatore India

ISBN: (数字)9798350387063

ISBN: (纸本)9798350387070

Autonomous navigation systems for robots are becoming increasingly essential in various domains, including transportation, surveillance, and logistics. Addressing the challenge of lane keeping is fundamental for safe and efficient navigation, mirroring the capabilities of human drivers. In this work, we present a camera-sensor based deep learning lane keeping system to autonomously move robots while guaranteeing they stay in defined lanes. The need for such a system arises from the rising demand for autonomous robotic solutions in real-world scenarios, where precise lane adherence is critical for safety and effectiveness. By leveraging convolutional neural networks (CNNs), our system accurately detects lane boundaries in real-time, ensuring robustness across diverse environmental conditions. A closed-loop control mechanism translates the CNN output into steering commands, facilitating smooth and precise navigation. Through extensive simulations and real-world experiments, we demonstrate the effectiveness and reliability of our approach, contributing to the advancement of autonomous robotic systems and enhancing their capability for safe and reliable navigation in various applications.

关键词： Training Navigation Operating systems Computer network reliability Surveillance Transportation Safety Convolutional neural networks Robots Autonomous robots

Paddy Plant Disease Detection using Hybrid Deep Learning Algorithms

学校读者我要写书评

暂无评论

Paddy Plant Disease Detection using Hybrid Deep Learning Alg...

International conference of Electronics, Communication and Aerospace Technology (ICECA)

作者： R Suguna S Deepa T Kanimozhi M Chandru G Arun Harshini M Department of Computer Science and Engineering Bannari Amman Institute of Technology Sathyamangalam Erode Tamilnadu India Department of Computer Technology - PG Kongu Engineering College Perundurai Erode Tamilnadu Department of Artificial Intelligence and Data Science Sri Eshwar College of Engineering Coimbatore Department of Computer Science VET-Institute of Arts and Science (Co-Education) College Erode Department of Computer Science and Engineering School of Computing Vel Tech Rangarajan Dr Sagunthala R&D Institute of Science and Technology Chennai M.Sc (Software Systems) Kongu Engineering College Perundurai Erode Tamilnadu

ISBN: (数字)9798350367904

ISBN: (纸本)9798350367911

Agriculture is the backbone of India’s economy, supporting a large portion of the population and establishing the country as a key global player in food production. India ranks as the second-largest producer of both paddy and wheat. However, agricultural productivity is highly susceptible to various plant diseases, which can significantly reduce crop yields and quality. Paddy, one of the most important staple crops, is particularly vulnerable to diseases caused mainly by viral and bacterial related diseases. The climate conditions are also play a vital role in paddy plant diseases. This research work proposes a deep learning-based approach for early detection and classification of paddy diseases. The proposed research devises a hybrid model that combines the VGG19 and ResNet50 algorithms, the system analyses images of both healthy and disease-affected paddy leaves, sourced from the Kaggle repository. This research focuses on predicting four major paddy diseases that can cause yield losses of up to 80%, and in severe cases, can destroy an entire crop: (i) Paddy Blast, (ii) Bacterial Leaf Blight, (iii) Sheath Blight, and (iv) Ufra Disease. The proposed model achieves a high accuracy rate of 98% in predicting paddy diseases and offers a rapid classification process, showcasing the potential of advanced image processing and deep learning techniques to mitigate the impact of paddy plant diseases on agricultural productivity.

关键词： Deep learning Productivity Plant diseases Microorganisms Accuracy Prediction algorithms Real-time systems Classification algorithms System analysis and design Residual neural networks

A Probabilistic Framework for Pruning Transformers Via a Finite Admixture of Keys 48

学校读者我要写书评

暂无评论

A Probabilistic Framework for Pruning Transformers Via a Fin...

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Nguyen, Tan M. Nguyen, Tam Bui, Long Do, Hai Nguyen, Duy Khuong Le, Dung D. Tran-The, Hung Ho, Nhat Osher, Stan J. Baraniuk, Richard G. University of California Department of Mathematics Los Angeles United States Rice University Department of Electrical and Computer Engineering Houston United States Fpt Software Ai Center Ha Noi Viet Nam Vin University College of Engineering and Computer Science Viet Nam Deakin University Applied Artificial Intelligence Institute Geelong Australia The University of Texas Department of Statistics and Data Sciences Austin United States

ISBN: (纸本)9781728163277

Pairwise dot product-based self-attention is key to the success of transformers which achieve state-of-the-art performance across a variety of applications in language and vision, but are costly to compute. It has been shown that most attention scores and keys in transformers are redundant and can be removed without loss of accuracy. In this paper, we develop a novel probabilistic framework for pruning attention scores and keys in transformers. We first formulate an admixture model of attention keys whose input data to be clustered are attention queries. We show that attention scores in self-attention correspond to the posterior distribution of this model when attention keys admit a uniform prior distribution. We then relax this uniform prior constraint and let the model learn these priors from data, resulting in a new Finite Admixture of Keys (FiAK). The learned priors are used for pruning away redundant attention scores and keys in the baseline transformers, improving the diversity of attention patterns that the models capture. We corroborate the efficiency of transformers pruned with FiAK on the ImageNet object classification and WikiText-103 language modeling tasks. Our experiments demonstrate that transformers pruned with FiAK yield similar or better accuracy than the baseline dense transformers while being much more efficient in terms of memory and computational cost. © 2023 IEEE.

关键词： admixture models pruning Transformers

Confidence-aware Contrastive Learning for Selective Classification

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wu, Yu-Chang Lyu, Shen-Huan Shang, Haopu Wang, Xiangyu Qian, Chao National Key Laboratory for Novel Software Technology Nanjing University China School of Artificial Intelligence Nanjing University China Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai University China College of Computer Science and Software Engineering Hohai University China

Selective classification enables models to make predictions only when they are sufficiently confident, aiming to enhance safety and reliability, which is important in high-stakes scenarios. Previous methods mainly use deep neural networks and focus on modifying the architecture of classification layers to enable the model to estimate the confidence of its prediction. This work provides a generalization bound for selective classification, disclosing that optimizing feature layers helps improve the performance of selective classification. Inspired by this theory, we propose to explicitly improve the selective classification model at the feature level for the first time, leading to a novel Confidence-aware Contrastive Learning method for Selective Classification, CCL-SC, which similarizes the features of homogeneous instances and differentiates the features of heterogeneous instances, with the strength controlled by the model’s confidence. The experimental results on typical datasets, i.e., CIFAR-10, CIFAR-100, CelebA, and ImageNet, show that CCL-SC achieves significantly lower selective risk than state-of-the-art methods, across almost all coverage degrees. Moreover, it can be combined with existing methods to bring further improvement. © 2024, CC BY.

关键词： Classification (of information)

The Role of Depth, Width, and Tree Size in Expressiveness of Deep Forest

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Lyu, Shen-Huan Wu, Jin-Hui Zheng, Qin-Cheng Ye, Baoliu Key Laboratory of Water Big Data Technology of Ministry of Water Resources Hohai University China College of Computer Science and Software Engineering Hohai University China National Key Laboratory for Novel Software Technology Nanjing University China School of Artificial Intelligence Nanjing University China

Random forests are classical ensemble algorithms that construct multiple randomized decision trees and aggregate their predictions using naive averaging. Zhou and Feng [50] further propose a deep forest algorithm with multi-layer forests, which outperforms random forests in various tasks. The performance of deep forests is related to three hyperparameters in practice: depth, width, and tree size, but little has been known about its theoretical explanation. This work provides the first upper and lower bounds on the approximation complexity of deep forests concerning the three hyperparameters. Our results confirm the distinctive role of depth, which can exponentially enhance the expressiveness of deep forests compared with width and tree size. Experiments confirm the theoretical findings. Copyright © 2024, The Authors. All rights reserved.

关键词： Decision trees

Searching for an Accurate Robot Calibration via Improved Levenberg–Marquardt and Radial Basis Function System

学校读者我要写书评

暂无评论

Journal of Field Robotics 2025年第0期

作者： Li, Zhibin Deng, Xun Chen, Tinghui Yang, Yuhang Chen, Linlin Yang, Xiwen Hu, Zhenzhen Hu, Lun Hu, Pengwei Li, Shuai Luo, Xin Xinjiang Technical Institute of Physics & Chemistry Chinese Academy of Sciences Urumqi China School of Software Engineering Chengdu University of Information Technology Chengdu China Dazhou Key Laboratory of Government Data Security Sichuan University of Arts and Science Dazhou China School of Software Engineering Xinjiang University Urumqi China Faculty of Information Technology and Electrical Engineering University of Oulu Oulu Finland VTT-Technical Research Centre of Finland Oulu Finland College of Computer and Information Science Southwest University Chongqing China

Robots are frequently utilized in manufacturing, aviation, and other industries, which enhance industrial production efficiency and quality. Specifically, robots perform high-precision tasks like welding, assembly and material handling, which reduce the intensity of manual labor in factories. In the logistics field, robots automatically sort and deliver goods, thereby speeding up supply chain operations. However, the prolonged operation of robots suffers from a decline in positioning accuracy, which makes them unable to satisfy task requirements. To address this challenging issue, this study designs an efficient calibration system integrating the Levenberg–Marquardt algorithm with fuzzy proportion integration differentiation controller and radial basis function neural network. The innovations of this method include: (1) integrating the fuzzy proportion integration differentiation controller into the updating rules of Levenberg–Marquardt algorithm, which further enhances the identification performance of kinematic errors;(2) adopting the radial basis function neural network to handle the robot dynamic errors, which addresses the complexity of dynamic error sources. Extensively experimental robot positioning points are gathered on an HSR JR680 robot, and then experimental validations are conducted by using the designed calibration system. The experiments indicate that the developed algorithm outperforms these existing advanced algorithms. © 2025 Wiley Periodicals LLC.

关键词： Radial basis function networks