Transfer_Learning
-
Automated Wicket-Taking Delivery Segmentation and Weakness Detection in Cricket Videos Using OCR-Guided YOLOv8 and Trajectory Modeling
2025-10-21 08:27:23Mst Jannatun Ferdous, Masum Billah, Joy Karmoker, Mohd Ruhul Ameen, Akif Islam, Md. Omar FaruqeAbstract
This paper presents an automated system for cricket video analysis that leverages deep learning techniques to extract wicket-taking deliveries, detect cricket balls, and model ball trajectories. The system employs the YOLOv8 architecture for pitch and ball detection, combined with optical character recognition (OCR) for scorecard extraction to identify wicket-taking moments. Through comprehensive image preprocessing, including grayscale transformation, power transformation, and morphological operations, the system achieves robust text extraction from video frames. The pitch detection model achieved 99.5% mean Average Precision at 50% IoU (mAP50) with a precision of 0.999, while the ball detection model using transfer learning attained 99.18% mAP50 with 0.968 precision and 0.978 recall. The system enables trajectory modeling on detected pitches, providing data-driven insights for identifying batting weaknesses. Experimental results on multiple cricket match videos demonstrate the effectiveness of this approach for automated cricket analytics, offering significant potential for coaching and strategic decision-making.
Abstract (translated)
这篇论文介绍了一个利用深度学习技术进行板球视频分析的自动化系统，该系统能够提取获得击球手出局时刻的关键投球、检测板球并建模其运动轨迹。系统使用YOLOv8架构结合光学字符识别（OCR）来实现场地和板球的检测，并从记分卡中提取信息以确定击球员出局的时间点。通过包括灰度转换、功率变换以及形态学操作在内的全面图像预处理，该系统能够有效地从视频帧中提取文本。 在球场检测模型方面，其精度达到了99.5%的平均精确率（mAP50），且IoU为50%，精度值为0.999。而在使用迁移学习的板球检测模型上，其精度则达到99.18%mAP50，具有0.968的精度和0.978的召回率。 该系统能够对已识别出的场地进行轨迹建模，从而为发现击球员的弱点提供了数据驱动的洞察。在多个板球比赛视频上的实验结果表明，这种方法对于自动化板球分析非常有效，并为教练工作以及战略决策提供重要的潜力。
URL
https://arxiv.org/abs/2510.18405
PDF
-
Lingua Custodi's participation at the WMT 2025 Terminology shared task
2025-10-20 13:00:47Jingshu Liu, Raheel Qader, Ga\"etan Caillaut, Mariam Nakhl\'eAbstract
While BERT is an effective method for learning monolingual sentence embeddings for semantic similarity and embedding based transfer learning BERT based cross-lingual sentence embeddings have yet to be explored. We systematically investigate methods for learning multilingual sentence embeddings by combining the best methods for learning monolingual and cross-lingual representations including: masked language modeling (MLM), translation language modeling (TLM), dual encoder translation ranking, and additive margin softmax. We show that introducing a pre-trained multilingual language model dramatically reduces the amount of parallel training data required to achieve good performance by 80%. Composing the best of these methods produces a model that achieves 83.7% bi-text retrieval accuracy over 112 languages on Tatoeba, well above the 65.5 achieved by LASER, while still performing competitively on monolingual transfer learning benchmarks. Parallel data mined from CommonCrawl using our best model is shown to train competitive NMT models for en-zh and en-de. We publicly release our best multilingual sentence embedding model for 109+ languages at this https URL.
Abstract (translated)
虽然BERT在学习单语句子嵌入以进行语义相似度和基于嵌入的迁移学习方面非常有效，但基于BERT的跨语言句子嵌入尚未得到充分研究。我们系统地调查了结合最佳单语文本表示与跨语言文本表示方法来学习多语言句子嵌入的方法，包括掩码语言模型（MLM）、翻译语言模型（TLM）、双编码器翻译排名以及加性余弦softmax。我们表明，引入预训练的多语种语言模型可以显著减少实现良好性能所需的平行训练数据量达80%。将这些方法中的最佳组合产生了一个在Tatoeba上涵盖112种语言且双文本检索准确率达到83.7%的模型，这一成绩远超LASER所达到的65.5%，同时该模型仍能在单语迁移学习基准测试中保持竞争力。使用我们的最佳模型从CommonCrawl中挖掘出的平行数据训练了竞争性的英汉和英德NMT（神经机器翻译）模型。我们公开发布了适用于109+种语言的最佳多语言句子嵌入模型，可在此URL下载：[https://github.com/your-repo-url](https://github.com/your-repo-url)。 请注意，在提供的链接处应填入实际的GitHub仓库地址或其他发布平台地址以供访问。
URL
https://arxiv.org/abs/2510.17504
PDF
-
Integrating Trustworthy Artificial Intelligence with Energy-Efficient Robotic Arms for Waste Sorting
2025-10-20 10:52:02Halima I. Kure, Jishna Retnakumari, Augustine O. Nwajana, Umar M. Ismail, Bilyaminu A. Romo, Ehigiator Egho-PromiseAbstract
This paper presents a novel methodology that integrates trustworthy artificial intelligence (AI) with an energy-efficient robotic arm for intelligent waste classification and sorting. By utilizing a convolutional neural network (CNN) enhanced through transfer learning with MobileNetV2, the system accurately classifies waste into six categories: plastic, glass, metal, paper, cardboard, and trash. The model achieved a high training accuracy of 99.8% and a validation accuracy of 80.5%, demonstrating strong learning and generalization. A robotic arm simulator is implemented to perform virtual sorting, calculating the energy cost for each action using Euclidean distance to ensure optimal and efficient movement. The framework incorporates key elements of trustworthy AI, such as transparency, robustness, fairness, and safety, making it a reliable and scalable solution for smart waste management systems in urban settings.
Abstract (translated)
本文提出了一种新颖的方法，该方法将值得信赖的人工智能（AI）与节能机器人手臂相结合，用于智能垃圾分类和分拣。通过利用基于MobileNetV2的迁移学习增强的卷积神经网络（CNN），系统能够准确地将垃圾分类为六类：塑料、玻璃、金属、纸张、硬纸板和其他垃圾。该模型在训练阶段达到了99.8%的高精度，在验证阶段则达到80.5%的准确性，显示出了强大的学习能力和泛化能力。 为了模拟虚拟分拣过程，设计了一个机器人手臂仿真器，通过计算每个动作的欧几里得距离来估算能耗，确保操作既高效又节能。该框架集成了值得信赖的人工智能的关键要素，包括透明性、鲁棒性、公平性和安全性，使其成为城市环境中智能垃圾分类系统可靠且可扩展的解决方案。
URL
https://arxiv.org/abs/2510.17408
PDF
-
Adaptive transfer learning for surgical tool presence detection in laparoscopic videos through gradual freezing fine-tuning
2025-10-17 07:17:52Ana Davila, Jacinto Colan, Yasuhisa HasegawaAbstract
Minimally invasive surgery can benefit significantly from automated surgical tool detection, enabling advanced analysis and assistance. However, the limited availability of annotated data in surgical settings poses a challenge for training robust deep learning models. This paper introduces a novel staged adaptive fine-tuning approach consisting of two steps: a linear probing stage to condition additional classification layers on a pre-trained CNN-based architecture and a gradual freezing stage to dynamically reduce the fine-tunable layers, aiming to regulate adaptation to the surgical domain. This strategy reduces network complexity and improves efficiency, requiring only a single training loop and eliminating the need for multiple iterations. We validated our method on the Cholec80 dataset, employing CNN architectures (ResNet-50 and DenseNet-121) pre-trained on ImageNet for detecting surgical tools in cholecystectomy endoscopic videos. Our results demonstrate that our method improves detection performance compared to existing approaches and established fine-tuning techniques, achieving a mean average precision (mAP) of 96.4%. To assess its broader applicability, the generalizability of the fine-tuning strategy was further confirmed on the CATARACTS dataset, a distinct domain of minimally invasive ophthalmic surgery. These findings suggest that gradual freezing fine-tuning is a promising technique for improving tool presence detection in diverse surgical procedures and may have broader applications in general image classification tasks.
Abstract (translated)
微创手术可以从自动化的外科工具检测中获益匪浅，这能够实现高级分析和辅助。然而，在手术环境中注释数据的有限可用性给训练健壮的深度学习模型带来了挑战。本文提出了一种新颖的分阶段自适应微调方法，该方法由两个步骤组成：线性探测阶段，用于在预训练的基于CNN（卷积神经网络）架构上附加分类层，并且一个渐进式冻结阶段，旨在动态减少可精调的层数，以调节对手术领域的适应。这种策略减少了网络复杂度并提高了效率，只需要一次训练循环即可完成，消除了多轮迭代的需求。 我们在Cholec80数据集上验证了我们的方法，该数据集中使用的是预训练于ImageNet上的CNN架构（如ResNet-50和DenseNet-121），用于在胆囊内镜视频中检测手术工具。我们的结果显示，相较于现有方法及传统的微调技术，本方法在检测性能方面有所提升，并且达到了96.4%的平均精度均值(mAP)。 为了评估其更广泛的应用性，我们还在CATARACTS数据集上进一步验证了微调策略的泛化能力。CATARACTS是一个完全不同的领域，涉及微创眼科手术。这些发现表明，渐进式冻结微调技术在不同类型的外科手术中对工具存在检测改进方面具有潜力，并可能在更广泛的图像分类任务中得到应用。
URL
https://arxiv.org/abs/2510.15372
PDF
-
Real-Time Knee Angle Prediction Using EMG and Kinematic Data with an Attention-Based CNN-LSTM Network and Transfer Learning Across Multiple Datasets
2025-10-15 11:41:40Mojtaba Mollahossein, Gholamreza Vossoughi, Mohammad Hossein RohbanAbstract
Electromyography (EMG) signals are widely used for predicting body joint angles through machine learning (ML) and deep learning (DL) methods. However, these approaches often face challenges such as limited real-time applicability, non-representative test conditions, and the need for large datasets to achieve optimal performance. This paper presents a transfer-learning framework for knee joint angle prediction that requires only a few gait cycles from new subjects. Three datasets - Georgia Tech, the University of California Irvine (UCI), and the Sharif Mechatronic Lab Exoskeleton (SMLE) - containing four EMG channels relevant to knee motion were utilized. A lightweight attention-based CNN-LSTM model was developed and pre-trained on the Georgia Tech dataset, then transferred to the UCI and SMLE datasets. The proposed model achieved Normalized Mean Absolute Errors (NMAE) of 6.8 percent and 13.7 percent for one-step and 50-step predictions on abnormal subjects using EMG inputs alone. Incorporating historical knee angles reduced the NMAE to 3.1 percent and 3.5 percent for normal subjects, and to 2.8 percent and 7.5 percent for abnormal subjects. When further adapted to the SMLE exoskeleton with EMG, kinematic, and interaction force inputs, the model achieved 1.09 percent and 3.1 percent NMAE for one- and 50-step predictions, respectively. These results demonstrate robust performance and strong generalization for both short- and long-term rehabilitation scenarios.
Abstract (translated)
肌电图（EMG）信号通过机器学习（ML）和深度学习（DL）方法被广泛用于预测人体关节角度。然而，这些方法通常面临诸如实时应用能力有限、测试条件代表性不足以及需要大量数据以达到最佳性能等问题。本文提出了一种转移学习框架，仅需新受试者几个步态周期的数据即可实现膝关节角度的预测。研究利用了三个数据集——佐治亚理工学院（Georgia Tech）、加州大学欧文分校（UCI）和夏里夫机械实验室外骨骼（SMLE），这些数据集中包含了与膝关节运动相关的四个EMG通道。 研究人员开发了一个轻量级基于注意力机制的CNN-LSTM模型，并在佐治亚理工大学的数据集上进行了预训练，然后将其转移到UCI和SMLE数据集。对于异常受试者，在仅使用EMG输入的情况下，该模型分别实现了6.8% 和13.7% 的归一化均方绝对误差（NMAE）的一步预测和50步预测。当引入历史膝关节角度信息后，对正常受试者的一步预测和50步预测的NMAE分别降至了3.1% 和3.5%，而对于异常受试者则分别为2.8% 和7.5%。 进一步将模型适应于SMLE外骨骼，并加入EMG、运动学以及交互力输入后，该模型对于一步预测和50步预测的NMAE分别达到了1.09% 和3.1%，展示了其在短期和长期康复场景中的强大泛化能力和稳健性能。
URL
https://arxiv.org/abs/2510.13443
PDF
-
TopoAlign: A Framework for Aligning Code to Math via Topological Decomposition
2025-10-13 21:09:36Yupei Li, Philipp Borchert, Gerasimos LampourasAbstract
Large Language Models (LLMs) excel at both informal and formal (e.g. Lean 4) mathematical reasoning but still struggle with autoformalisation, the task of transforming informal into formal mathematical statements. Autoformalisation helps pair the informal reasoning of LLMs with formal proof assistants which enable machine-verifiable generation and mitigate hallucinations. Yet, the performance of current Math LLMs is constrained by the scarcity of large-scale corpora, particularly those containing pairs of informal and formal statements. Although current models are trained to generate code from natural language instructions, structural and syntactic differences between these and formal mathematics limit effective transfer learning. We propose TopoAlign, a framework that unlocks widely available code repositories as training resources for Math LLMs. TopoAlign decomposes code into docstrings, main functions, and dependency functions, and reassembles these components into analogues that structurally mirror formal statements. This produces structurally aligned code data that can be used for training Math LLMs without requiring additional human annotation. We train two state-of-the-art models, DeepSeek-Math and Herald, and evaluate them on the minif2f, Putnam, and ProofNet benchmarks. TopoAlign provides substantial gains for DeepSeek-Math, improving performance by 17.77% on BEq@10 and 68.82% on typecheck@10. Despite introducing no new mathematical knowledge, our framework achieves gains of 0.12% and 1.09% for Herald on BEq@10 and typecheck@10, respectively, demonstrating that training on aligned code data is beneficial even for specialized models.
Abstract (translated)
大型语言模型（LLMs）在非正式和正式数学推理（如Lean 4）方面表现出色，但它们仍然在自动形式化任务上遇到困难。自动形式化的目的是将非正式的数学陈述转换为正式的数学表述，这一过程有助于将LLM的非正式推理与证明辅助工具相匹配，从而实现机器可验证的生成并减少幻觉（错误信息）。然而，当前数学模型的表现受限于大规模语料库的稀缺性，尤其是包含非正式和正式声明配对的语料库。尽管现有的模型被训练从自然语言指令生成代码，但这些语言与正式数学之间的结构和语法差异限制了有效的迁移学习。 我们提出了一个名为TopoAlign的框架，该框架解锁了广泛可用的代码仓库作为数学LLM的培训资源。TopoAlign将代码分解为文档字符串、主函数和依赖函数，并重新组合成类似于正式声明的结构镜像体，从而生成可以用于训练数学模型且无需额外人工注释的结构对齐代码数据。 我们使用两个最先进的模型——DeepSeek-Math和Herald，在minif2f、Putnam和ProofNet基准测试上进行实验。TopoAlign框架为DeepSeek-Math带来了显著改善，在BEq@10上的性能提升了17.77%，在typecheck@10上的性能提高了68.82%。尽管我们的框架并未引入新的数学知识，但仍然使得Herald模型的BEq@10和typecheck@10分别获得了0.12%和1.09%的改进，这表明即使是专门化的模型也能够从对齐后的代码数据训练中受益。 该研究证明了通过结构化映射现有代码资源以适应数学LLM培训的有效性，并展示了利用这种资源改善模型性能的巨大潜力。
URL
https://arxiv.org/abs/2510.11944
PDF
-
Nepali Sign Language Characters Recognition: Dataset Development and Deep Learning Approaches
2025-10-13 10:29:08Birat Poudel, Satyam Ghimire, Sijan Bhattarai, Saurav Bhandari, Suramya Sharma DahalAbstract
Sign languages serve as essential communication systems for individuals with hearing and speech impairments. However, digital linguistic dataset resources for underrepresented sign languages, such as Nepali Sign Language (NSL), remain scarce. This study introduces the first benchmark dataset for NSL, consisting of 36 gesture classes with 1,500 samples per class, designed to capture the structural and visual features of the language. To evaluate recognition performance, we fine-tuned MobileNetV2 and ResNet50 architectures on the dataset, achieving classification accuracies of 90.45% and 88.78%, respectively. These findings demonstrate the effectiveness of convolutional neural networks in sign recognition tasks, particularly within low-resource settings. To the best of our knowledge, this work represents the first systematic effort to construct a benchmark dataset and assess deep learning approaches for NSL recognition, highlighting the potential of transfer learning and fine-tuning for advancing research in underexplored sign languages.
Abstract (translated)
手语作为听力和言语障碍人士的重要沟通方式，发挥着不可或缺的作用。然而，对于如尼泊尔手语（NSL）等代表性不足的手语来说，数字语言数据集资源仍然非常稀缺。本研究首次介绍了针对NSL的基准数据集，该数据集包含36个手势类别，每个类别有1500个样本，旨在捕捉该语言的结构和视觉特征。 为了评估识别性能，我们在MobileNetV2和ResNet50架构上对数据集进行了微调，分别实现了90.45%和88.78%的分类准确率。这些发现证明了卷积神经网络在手语识别任务中的有效性，尤其是在资源匮乏的情况下。 据我们所知，这项工作代表了首次系统性地为NSL构建基准数据集并评估深度学习方法的研究尝试，突显了迁移学习和微调技术对于探索未被充分研究的手语领域的潜力。
URL
https://arxiv.org/abs/2510.11243
PDF
-
Image-to-Video Transfer Learning based on Image-Language Foundation Models: A Comprehensive Survey
2025-10-12 15:56:02Jinxuan Li, Chaolei Tan, Haoxuan Chen, Jianxin Ma, Jian-Fang Hu, Wei-Shi Zheng, Jianhuang LaiAbstract
Image-Language Foundation Models (ILFM) have demonstrated remarkable success in image-text understanding/generation tasks, providing transferable multimodal representations that generalize across diverse downstream image-based tasks. The advancement of video-text research has spurred growing interest in extending image-based models to the video domain. This paradigm, known as image-to-video transfer learning, succeeds in alleviating the substantial data and computational requirements associated with training video-language foundation models from scratch for video-text learning. This survey provides the first comprehensive review of this emerging field, which begins by summarizing the widely used ILFM and their capabilities. We then systematically classify existing image-to-video transfer learning strategies into two categories: frozen features and modified features, depending on whether the original representations from ILFM are preserved or undergo modifications. Building upon the task-specific nature of image-to-video transfer, this survey methodically elaborates these strategies and details their applications across a spectrum of video-text learning tasks, ranging from fine-grained (e.g., spatio-temporal video grounding) to coarse-grained (e.g., video question answering). We further present a detailed experimental analysis to investigate the efficacy of different image-to-video transfer learning paradigms on a range of downstream video understanding tasks. Finally, we identify prevailing challenges and highlight promising directions for future research. By offering a comprehensive and structured overview, this survey aims to establish a structured roadmap for advancing video-text learning based on existing ILFM, and to inspire future research directions in this rapidly evolving domain.
Abstract (translated)
图像-语言基础模型（ILFM）在图像文本理解和生成任务中展示了显著的成功，提供了可转移的多模态表示形式，这些表示能够在各种下游基于图像的任务中进行泛化。视频文本研究的进步激发了将基于图像的模型扩展到视频领域的日益增长的兴趣。这一范式被称为从图像到视频的迁移学习，在缓解训练用于视频文本学习的视频语言基础模型所需的大量数据和计算资源需求方面取得了成功。 这篇综述提供了对该新兴领域首次全面的回顾，首先总结了广泛使用的ILFM及其能力。然后，我们系统地将现有的图像到视频的迁移学习策略分为两类：冻结特征和修改特征，这取决于原始ILFM表示是否被保留或进行修改。基于从图像到视频迁移的任务特定性质，该综述详细阐述这些策略，并介绍了它们在一系列视频文本学习任务中的应用，从细粒度（例如时空视频定位）到粗粒度（例如视频问答）。我们进一步进行了详细的实验分析，以研究不同图像到视频的迁移学习范式对各种下游视频理解任务的有效性。最后，我们识别了当前存在的挑战，并指出了未来研究中具有前景的方向。 通过提供全面和结构化的概述，本综述旨在为基于现有ILFM的视频文本学习建立一个有条理的发展路线图，并激发这一迅速发展的领域中的未来研究方向。
URL
https://arxiv.org/abs/2510.10671
PDF
-
Self-Attention to Operator Learning-based 3D-IC Thermal Simulation
2025-10-12 13:44:42Zhen Huang, Hong Wang, Wenkai Yang, Muxi Tang, Depeng Xie, Ting-Jung Lin, Yu Zhang, Wei W. Xing, Lei HeAbstract
Thermal management in 3D ICs is increasingly challenging due to higher power densities. Traditional PDE-solving-based methods, while accurate, are too slow for iterative design. Machine learning approaches like FNO provide faster alternatives but suffer from high-frequency information loss and high-fidelity data dependency. We introduce Self-Attention U-Net Fourier Neural Operator (SAU-FNO), a novel framework combining self-attention and U-Net with FNO to capture long-range dependencies and model local high-frequency features effectively. Transfer learning is employed to fine-tune low-fidelity data, minimizing the need for extensive high-fidelity datasets and speeding up training. Experiments demonstrate that SAU-FNO achieves state-of-the-art thermal prediction accuracy and provides an 842x speedup over traditional FEM methods, making it an efficient tool for advanced 3D IC thermal simulations.
Abstract (translated)
在三维集成电路（3D ICs）中，热管理由于更高的功率密度而变得越来越具有挑战性。传统的基于偏微分方程(PDE)求解的方法虽然准确，但速度太慢，不适合迭代设计过程。机器学习方法如傅里叶神经网络(FNO)提供了更快的替代方案，但它们在高频信息丢失和高保真数据依赖方面存在问题。 我们引入了一种新的框架——Self-Attention U-Net Fourier Neural Operator (SAU-FNO)，该框架结合了自注意力机制、U-Net架构与FNO，可以有效捕捉长距离依赖性并准确建模局部高频特征。同时采用了迁移学习来对低保真数据进行微调，从而减少了对大规模高保真数据集的需求，并加快了训练过程。 实验结果显示，SAU-FNO在热预测准确性上达到了业界领先水平，并且比传统的有限元方法(FEM)快842倍，成为先进3D IC热模拟中一种高效工具。
URL
https://arxiv.org/abs/2510.15968
PDF
-
Towards Cybersickness Severity Classification from VR Gameplay Videos Using Transfer Learning and Temporal Modeling
2025-10-12 03:10:05Jyotirmay Nag Setu, Kevin Desai, John QuarlesAbstract
With the rapid advancement of virtual reality (VR) technology, its adoption across domains such as healthcare, education, and entertainment has grown significantly. However, the persistent issue of cybersickness, marked by symptoms resembling motion sickness, continues to hinder widespread acceptance of VR. While recent research has explored multimodal deep learning approaches leveraging data from integrated VR sensors like eye and head tracking, there remains limited investigation into the use of video-based features for predicting cybersickness. In this study, we address this gap by utilizing transfer learning to extract high-level visual features from VR gameplay videos using the InceptionV3 model pretrained on the ImageNet dataset. These features are then passed to a Long Short-Term Memory (LSTM) network to capture the temporal dynamics of the VR experience and predict cybersickness severity over time. Our approach effectively leverages the time-series nature of video data, achieving a 68.4% classification accuracy for cybersickness severity. This surpasses the performance of existing models trained solely on video data, providing a practical tool for VR developers to evaluate and mitigate cybersickness in virtual environments. Furthermore, this work lays the foundation for future research on video-based temporal modeling for enhancing user comfort in VR applications.
Abstract (translated)
随着虚拟现实（VR）技术的迅速发展，它在医疗保健、教育和娱乐等领域的应用显著增加。然而，持续存在的网络眩晕问题——其症状类似于运动病——继续阻碍着VR的广泛接受度。尽管最近的研究探讨了多模态深度学习方法，这些方法利用集成的VR传感器（如眼睛跟踪和头部跟踪）的数据，但对于基于视频特征预测网络眩晕的应用研究仍较为有限。 在这项研究中，我们填补了这一空白，通过使用InceptionV3模型从预训练于ImageNet数据集的VR游戏视频中提取高级视觉特征，并采用迁移学习技术。然后，这些特征被传递给长短期记忆（LSTM）网络以捕捉VR体验的时间动态性并预测随着时间变化的网络眩晕严重程度。 我们的方法有效地利用了视频数据的时间序列特性，在网络眩晕严重度分类上达到了68.4%的准确率。这超过了仅基于视频数据训练的现有模型的表现，为VR开发者提供了一种实用工具，用于评估和缓解虚拟环境中的网络眩晕问题。此外，这项工作还为未来研究打下了基础，即通过基于视频的时间建模来增强VR应用中用户的舒适度。
URL
https://arxiv.org/abs/2510.10422
PDF
-
deep-REMAP: Probabilistic Parameterization of Stellar Spectra Using Regularized Multi-Task Learning
2025-10-10 13:20:06Sankalp GildaAbstract
In the era of exploding survey volumes, traditional methods of spectroscopic analysis are being pushed to their limits. In response, we develop deep-REMAP, a novel deep learning framework that utilizes a regularized, multi-task approach to predict stellar atmospheric parameters from observed spectra. We train a deep convolutional neural network on the PHOENIX synthetic spectral library and use transfer learning to fine-tune the model on a small subset of observed FGK dwarf spectra from the MARVELS survey. We then apply the model to 732 uncharacterized FGK giant candidates from the same survey. When validated on 30 MARVELS calibration stars, deep-REMAP accurately recovers the effective temperature ($T_{\rm{eff}}$), surface gravity ($\log \rm{g}$), and metallicity ([Fe/H]), achieving a precision of, for instance, approximately 75 K in $T_{\rm{eff}}$. By combining an asymmetric loss function with an embedding loss, our regression-as-classification framework is interpretable, robust to parameter imbalances, and capable of capturing non-Gaussian uncertainties. While developed for MARVELS, the deep-REMAP framework is extensible to other surveys and synthetic libraries, demonstrating a powerful and automated pathway for stellar characterization.
Abstract (translated)
在天文观测数据量爆炸增长的时代，传统的光谱分析方法正面临极限挑战。为此，我们开发了深度学习框架deep-REMAP，该框架采用了一种正则化、多任务的方法来预测恒星大气参数（如有效温度$T_{\rm{eff}}$、表面重力$\log \rm{g}$和金属丰度[Fe/H]）从观测光谱中提取的信息。我们首先在PHOENIX合成光谱库上训练了一个深度卷积神经网络，然后利用迁移学习对来自MARVELS调查的小规模观测FGK矮星光谱数据进行微调。接着我们将模型应用于同一调查中的732个未表征的FGK巨星候选体。 当该框架在30颗MARVELS校准恒星上验证时，deep-REMAP能够准确地恢复有效温度（$T_{\rm{eff}}$）、表面重力（$\log \rm{g}$）和金属丰度([Fe/H])等参数，并且在$T_{\rm{eff}}$方面达到了大约75K的精度。通过结合非对称损失函数与嵌入式损失，我们的回归-分类框架具有可解释性、能够应对参数不平衡的问题，并能捕捉到非高斯不确定性。 尽管deep-REMAP最初是为MARVELS调查设计的，但该框架可以扩展至其他天文调查和合成光谱库中使用，展示了恒星表征领域的强大且自动化的方法路径。
URL
https://arxiv.org/abs/2510.09362
PDF
-
Rewiring Development in Brain Segmentation: Leveraging Adult Brain Priors for Enhancing Infant MRI Segmentation
2025-10-10 11:55:43Alemu Sisay Nigru, Michele Svanera, Austin Dibble, Connor Dalby, Mattia Savardi, Sergio BeniniAbstract
Accurate segmentation of infant brain MRI is critical for studying early neurodevelopment and diagnosing neurological disorders. Yet, it remains a fundamental challenge due to continuously evolving anatomy of the subjects, motion artifacts, and the scarcity of high-quality labeled data. In this work, we present LODi, a novel framework that utilizes prior knowledge from an adult brain MRI segmentation model to enhance the segmentation performance of infant scans. Given the abundance of publicly available adult brain MRI data, we pre-train a segmentation model on a large adult dataset as a starting point. Through transfer learning and domain adaptation strategies, we progressively adapt the model to the 0-2 year-old population, enabling it to account for the anatomical and imaging variability typical of infant scans. The adaptation of the adult model is carried out using weakly supervised learning on infant brain scans, leveraging silver-standard ground truth labels obtained with FreeSurfer. By introducing a novel training strategy that integrates hierarchical feature refinement and multi-level consistency constraints, our method enables fast, accurate, age-adaptive segmentation, while mitigating scanner and site-specific biases. Extensive experiments on both internal and external datasets demonstrate the superiority of our approach over traditional supervised learning and domain-specific models. Our findings highlight the advantage of leveraging adult brain priors as a foundation for age-flexible neuroimaging analysis, paving the way for more reliable and generalizable brain MRI segmentation across the lifespan.
Abstract (translated)
婴儿脑部MRI的精确分割对于研究早期神经发育和诊断神经系统疾病至关重要。然而，由于受试者解剖结构不断变化、运动伪影以及高质量标注数据稀缺等因素，这一任务仍面临重大挑战。在这项工作中，我们提出了LODi框架，利用成人脑部MRI分割模型中的先验知识来提高婴儿扫描的分割性能。鉴于公开可用的成人脑部MRI数据丰富，我们首先在一个大型成人数据集上预训练了一个分割模型作为起点。通过迁移学习和领域适应策略，逐步将该模型调整适用于0至2岁儿童群体，使其能够应对婴儿扫描中常见的解剖学和成像变异。 我们将成年人模型应用到婴儿脑部扫描的过程中采用了弱监督学习，并利用FreeSurfer获得的银标准地面真实标签进行训练。通过引入一种结合层级特征精炼和多级一致性约束的新颖训练策略，我们的方法实现了快速、精确且适应年龄变化的分割效果，同时减少了特定扫描仪和地点带来的偏差。 在内部和外部数据集上的广泛实验表明，相较于传统的监督学习和领域特化模型，我们所提出的方法表现更优。我们的发现强调了利用成人脑部先验知识作为跨年龄段灵活神经影像分析基础的重要性，并为整个生命周期内更加可靠且通用的脑部MRI分割铺平了道路。
URL
https://arxiv.org/abs/2510.09306
PDF
-
Modern Deep Learning Approaches for Cricket Shot Classification: A Comprehensive Baseline Study
2025-10-10 09:32:29Sungwoo KangAbstract
Cricket shot classification from video sequences remains a challenging problem in sports video analysis, requiring effective modeling of both spatial and temporal features. This paper presents the first comprehensive baseline study comparing seven different deep learning approaches across four distinct research paradigms for cricket shot classification. We implement and systematically evaluate traditional CNN-LSTM architectures, attention-based models, vision transformers, transfer learning approaches, and modern EfficientNet-GRU combinations on a unified benchmark. A critical finding of our study is the significant performance gap between claims in academic literature and practical implementation results. While previous papers reported accuracies of 96\% (Balaji LRCN), 99.2\% (IJERCSE), and 93\% (Sensors), our standardized re-implementations achieve 46.0\%, 55.6\%, and 57.7\% respectively. Our modern SOTA approach, combining EfficientNet-B0 with a GRU-based temporal model, achieves 92.25\% accuracy, demonstrating that substantial improvements are possible with modern architectures and systematic optimization. All implementations follow modern MLOps practices with PyTorch Lightning, providing a reproducible research platform that exposes the critical importance of standardized evaluation protocols in sports video analysis research.
Abstract (translated)
从视频序列中对板球击球动作进行分类仍然是体育视频分析中的一个挑战性问题，需要有效地建模空间和时间特征。本文首次进行了全面的基准研究，比较了七种不同的深度学习方法，在四个不同的研究范式下用于板球击球动作分类。我们实现了并系统地评估了几种传统CNN-LSTM架构、基于注意力的模型、视觉变换器、迁移学习方法以及现代EfficientNet-GRU组合在统一基准上的表现。 我们的研究中一个关键发现是学术文献中的报告与实际实现结果之间存在显著差异：尽管此前有论文报道了高达96%（Balaji LRCN）、99.2%（IJERCSE）和93%（Sensors）的准确率，我们统一标准重新实现的结果分别为46.0%，55.6% 和57.7%。我们的现代最先进方法，将EfficientNet-B0与基于GRU的时间模型结合使用，在准确性上达到了92.25%，这表明通过采用现代架构和系统的优化可以取得显著改进。 所有实现都遵循了最新的MLOps实践，并采用了PyTorch Lightning框架，为可重复研究提供了一个平台，突显了在体育视频分析研究中标准化评估协议的重要性。
URL
https://arxiv.org/abs/2510.09187
PDF
-
A Novel Multi-branch ConvNeXt Architecture for Identifying Subtle Pathological Features in CT Scans
2025-10-10 08:00:46Irash Perera (Department of Computer Science and Engineering, University of Moratuwa, Colombo, Sri Lanka), Uthayasanker Thayasivam (Department of Computer Science and Engineering, University of Moratuwa, Colombo, Sri Lanka)Abstract
Intelligent analysis of medical imaging plays a crucial role in assisting clinical diagnosis, especially for identifying subtle pathological features. This paper introduces a novel multi-branch ConvNeXt architecture designed specifically for the nuanced challenges of medical image analysis. While applied here to the specific problem of COVID-19 diagnosis, the methodology offers a generalizable framework for classifying a wide range of pathologies from CT scans. The proposed model incorporates a rigorous end-to-end pipeline, from meticulous data preprocessing and augmentation to a disciplined two-phase training strategy that leverages transfer learning effectively. The architecture uniquely integrates features extracted from three parallel branches: Global Average Pooling, Global Max Pooling, and a new Attention-weighted Pooling mechanism. The model was trained and validated on a combined dataset of 2,609 CT slices derived from two distinct datasets. Experimental results demonstrate a superior performance on the validation set, achieving a final ROC-AUC of 0.9937, a validation accuracy of 0.9757, and an F1-score of 0.9825 for COVID-19 cases, outperforming all previously reported models on this dataset. These findings indicate that a modern, multi-branch architecture, coupled with careful data handling, can achieve performance comparable to or exceeding contemporary state-of-the-art models, thereby proving the efficacy of advanced deep learning techniques for robust medical diagnostics.
Abstract (translated)
智能分析医学影像在辅助临床诊断中扮演着关键角色，特别是在识别细微的病理特征方面。本文介绍了一种新颖的多分支ConvNeXt架构，该架构专门针对医疗图像分析中的复杂挑战而设计。虽然这项研究主要应用于新冠肺炎（COVID-19）诊断这一具体问题上，但所提出的方法提供了一个通用框架，可用于从CT扫描中分类各种病理类型。提出的模型包括一个严格端到端的流程，涵盖了细致的数据预处理和增强以及一种严谨的两阶段训练策略，该策略有效利用了迁移学习的优势。该架构独特地整合了三种平行分支提取的功能：全局平均池化、全局最大池化及一个新的注意力加权池化机制。模型是在由两个不同数据集合并而成的包含2,609个CT切片的数据集上进行训练和验证的。实验结果表明，在验证集中取得了卓越的表现，最终达到了ROC-AUC为0.9937、验证准确率为0.9757及COVID-19病例F1分数为0.9825的结果，超过了在此数据集上的所有先前报告模型的表现。这些发现表明，现代多分支架构结合仔细的数据处理可以达到与当代最先进的模型相媲美或超过的性能水平，从而证明了高级深度学习技术在医学诊断中的有效性。
URL
https://arxiv.org/abs/2510.09107
PDF
-
Exploring Cross-Lingual Knowledge Transfer via Transliteration-Based MLM Fine-Tuning for Critically Low-resource Chakma Language
2025-10-10 06:07:14Adity Khisa, Nusrat Jahan Lia, Tasnim Mahfuz Nafis, Zarif Masud, Tanzir Pial, Shebuti Rayana, Ahmedul KabirAbstract
As an Indo-Aryan language with limited available data, Chakma remains largely underrepresented in language models. In this work, we introduce a novel corpus of contextually coherent Bangla-transliterated Chakma, curated from Chakma literature, and validated by native speakers. Using this dataset, we fine-tune six encoder-based multilingual and regional transformer models (mBERT, XLM-RoBERTa, DistilBERT, DeBERTaV3, BanglaBERT, and IndicBERT) on masked language modeling (MLM) tasks. Our experiments show that fine-tuned multilingual models outperform their pre-trained counterparts when adapted to Bangla-transliterated Chakma, achieving up to 73.54% token accuracy and a perplexity as low as 2.90. Our analysis further highlights the impact of data quality on model performance and shows the limitations of OCR pipelines for morphologically rich Indic scripts. Our research demonstrates that Bangla-transliterated Chakma can be very effective for transfer learning for Chakma language, and we release our manually validated monolingual dataset to encourage further research on multilingual language modeling for low-resource languages.
Abstract (translated)
作为一种印地语-雅利安语言，由于可用数据有限，查克马语在语言模型中严重代表性不足。在这项工作中，我们引入了一种新型的连贯上下文下的孟加拉语转写的查克马语语料库，该语料库由查克马文学作品整理而来，并经母语者验证确认。利用这个数据集，我们在六种基于编码器的多语言和区域变压器模型（mBERT、XLM-RoBERTa、DistilBERT、DeBERTaV3、孟加拉语BERT和印地语BERT）上进行了掩码语言建模（MLM）任务的微调。实验结果表明，当这些多语言模型适应于孟加拉语转写的查克马语时，它们的表现优于其预训练版本，达到了高达73.54%的标记准确率和低至2.90的困惑度。我们的分析进一步强调了数据质量对模型性能的影响，并揭示了光学字符识别（OCR）流程在形态丰富的印地语脚本中的局限性。我们的研究证明了孟加拉语转写的查克马语可以非常有效地用于查克马语言的迁移学习，我们发布了经过人工验证的单语文本数据集以促进对低资源语言多语言建模的研究。
URL
https://arxiv.org/abs/2510.09032
PDF
-
Denoised Diffusion for Object-Focused Image Augmentation
2025-10-10 03:03:40Nisha Pillai, Aditi Virupakshaiah, Harrison W. Smith, Amanda J. Ashworth, Prasanna Gowda, Phillip R. Owens, Adam R. Rivers, Bindu Nanduri, Mahalingam RamkumarAbstract
Modern agricultural operations increasingly rely on integrated monitoring systems that combine multiple data sources for farm optimization. Aerial drone-based animal health monitoring serves as a key component but faces limited data availability, compounded by scene-specific issues such as small, occluded, or partially visible animals. Transfer learning approaches often fail to address this limitation due to the unavailability of large datasets that reflect specific farm conditions, including variations in animal breeds, environments, and behaviors. Therefore, there is a need for developing a problem-specific, animal-focused data augmentation strategy tailored to these unique challenges. To address this gap, we propose an object-focused data augmentation framework designed explicitly for animal health monitoring in constrained data settings. Our approach segments animals from backgrounds and augments them through transformations and diffusion-based synthesis to create realistic, diverse scenes that enhance animal detection and monitoring performance. Our initial experiments demonstrate that our augmented dataset yields superior performance compared to our baseline models on the animal detection task. By generating domain-specific data, our method empowers real-time animal health monitoring solutions even in data-scarce scenarios, bridging the gap between limited data and practical applicability.
Abstract (translated)
现代农业操作越来越依赖于结合多个数据来源的综合监测系统，以实现农场优化。基于无人机的动物健康监测是这一过程中的关键组成部分，但面临着有限的数据可用性问题，尤其是在特定场景中，如小型、被遮挡或部分可见的动物情况下尤为突出。由于缺乏能够反映具体农场条件的大规模数据集（包括不同品种、环境和行为的变化），迁移学习方法往往无法有效解决这些问题。因此，有必要开发一种专门针对这些独特挑战的问题特定、以动物为中心的数据增强策略。 为了解决这一缺口，我们提出了一种面向对象的、专为在受限数据环境中进行动物健康监测设计的数据增强框架。我们的方法通过分割动物与背景，并通过对动物应用变换和基于扩散的合成来生成现实且多样的场景，从而增强动物检测和监控性能。初步实验表明，在动物检测任务上，使用我们扩增后的数据集比基线模型表现出更优的效果。 通过生成特定领域的数据，这种方法能够在数据匮乏的情况下为实时动物健康监测解决方案提供支持，并弥合了有限数据与实际应用之间的差距。
URL
https://arxiv.org/abs/2510.08955
PDF
-
Structured Output Regularization: a framework for few-shot transfer learning
2025-10-09 18:34:22Nicolas Ewen, Jairo Diaz-Rodriguez, Kelly RamsayAbstract
Traditional transfer learning typically reuses large pre-trained networks by freezing some of their weights and adding task-specific layers. While this approach is computationally efficient, it limits the model's ability to adapt to domain-specific features and can still lead to overfitting with very limited data. To address these limitations, we propose Structured Output Regularization (SOR), a simple yet effective framework that freezes the internal network structures (e.g., convolutional filters) while using a combination of group lasso and $L_1$ penalties. This framework tailors the model to specific data with minimal additional parameters and is easily applicable to various network components, such as convolutional filters or various blocks in neural networks enabling broad applicability for transfer learning tasks. We evaluate SOR on three few shot medical imaging classification tasks and we achieve competitive results using DenseNet121, and EfficientNetB4 bases compared to established benchmarks.
Abstract (translated)
传统的迁移学习通常通过冻结部分预训练网络的权重并添加特定任务层来重用大型预训练网络。虽然这种方法在计算上是高效的，但它限制了模型适应领域特有特征的能力，并且在数据非常有限的情况下仍可能导致过拟合。为了解决这些局限性，我们提出了一种名为结构化输出正则化（SOR）的简单而有效的框架，在该框架中冻结内部网络结构（例如卷积滤波器），同时使用组套索和$L_1$惩罚的组合方法。此框架能够利用最少的额外参数使模型适应特定数据，并且易于应用于各种网络组件，如卷积滤波器或神经网络中的各个块，从而在迁移学习任务中具有广泛的适用性。我们在三个少量样本医学图像分类任务上评估了SOR，并使用DenseNet121和EfficientNetB4作为基础模型时取得了与现有基准相当的结果。
URL
https://arxiv.org/abs/2510.08728
PDF
-
Trajectory Conditioned Cross-embodiment Skill Transfer
2025-10-09 04:26:06YuHang Tang, Yixuan Lou, Pengfei Han, Haoming Song, Xinyi Ye, Dong Wang, Bin ZhaoAbstract
Learning manipulation skills from human demonstration videos presents a promising yet challenging problem, primarily due to the significant embodiment gap between human body and robot manipulators. Existing methods rely on paired datasets or hand-crafted rewards, which limit scalability and generalization. We propose TrajSkill, a framework for Trajectory Conditioned Cross-embodiment Skill Transfer, enabling robots to acquire manipulation skills directly from human demonstration videos. Our key insight is to represent human motions as sparse optical flow trajectories, which serve as embodiment-agnostic motion cues by removing morphological variations while preserving essential dynamics. Conditioned on these trajectories together with visual and textual inputs, TrajSkill jointly synthesizes temporally consistent robot manipulation videos and translates them into executable actions, thereby achieving cross-embodiment skill transfer. Extensive experiments are conducted, and the results on simulation data (MetaWorld) show that TrajSkill reduces FVD by 39.6\% and KVD by 36.6\% compared with the state-of-the-art, and improves cross-embodiment success rate by up to 16.7\%. Real-robot experiments in kitchen manipulation tasks further validate the effectiveness of our approach, demonstrating practical human-to-robot skill transfer across embodiments.
Abstract (translated)
从人类演示视频中学习操作技能是一个有前景但又极具挑战性的问题，主要原因是人手和机器人执行器之间存在显著的身体差异。现有的方法依赖于成对的数据集或手工制作的奖励机制，这限制了可扩展性和泛化能力。我们提出了一个名为TrajSkill的框架，这是一种基于轨迹条件跨体形技能转移的方法，能够使机器人直接从人类演示视频中学习操作技能。 我们的核心观点是将人类的动作表示为稀疏光流轨迹，这些轨迹通过消除形态变化而保留了关键的动力学特征，并作为与具体身体形式无关的动作线索。在这些轨迹、视觉和文本输入的条件下，TrajSkill可以同时生成时间上一致的机器人操作视频，并将其转化为可执行的操作动作，从而实现跨体形技能转移。 我们进行了广泛的实验，在模拟数据（MetaWorld）上的结果显示，TrajSkill相比现有最佳方法减少了39.6%的FVD分数和36.6%的KVD分数，并将跨身体形式的成功率提高了高达16.7%。在厨房操作任务中的真实机器人实验进一步验证了我们方法的有效性，展示了从人类到机器人的实际技能转移能力，能够跨越不同的身体形式。
URL
https://arxiv.org/abs/2510.07773
PDF
-
Deploying Tiny LVLM Judges for Real-World Evaluation of Chart Models: Lessons Learned and Best Practices
2025-10-08 21:02:19Md Tahmid Rahman Laskar, Mohammed Saidul Islam, Ridwan Mahbub, Mizanur Rahman, Amran Bhuiyan, Israt Jahan, Mir Tafseer Nayeem, Shafiq Joty, Enamul Hoque, Jimmy HuangAbstract
Large Vision-Language Models (LVLMs) with only 7B parameters have shown promise as automated judges in chart comprehension tasks. However, tiny models (<=2B parameters) still perform poorly as judges, limiting their real-world use in resource-constrained settings. To address this, we propose two approaches to ensure cost-efficient evaluation: (i) multi-criteria prompting, which combines separate evaluation criteria into a single query, and (ii) domain-adaptive transfer learning, in which we fine-tune a 2B-parameter LVLM on synthetic judgments in a chart dataset to create the ChartJudge. Experiments show that multi-criteria prompting exposes robustness gaps, which led to a huge drop in performance for 7B models, including specialized LVLM judges like LLaVA-Critic. In addition, we find that our tiny LVLM (ChartJudge) can effectively transfer knowledge from one dataset to another to make it a more specialized model. Our fine-grained analysis across chart types and query complexities offers actionable insights into trade-offs between model size, prompt design, and transferability, enabling scalable, low-cost evaluation for chart reasoning tasks. Our code and the data will be made publicly available.
Abstract (translated)
具有70亿参数的大规模视觉-语言模型（LVLMs）在图表理解任务中作为自动评判工具方面显示出了潜力。然而，小型模型（<=20亿参数）在评判任务中的表现仍然不佳，限制了它们在资源受限环境下的实际应用。为了解决这一问题，我们提出了两种确保成本效益评估的方法：(i) 多标准提示法，这种方法将不同的评价标准合并到一个查询中；以及(ii) 领域适应性迁移学习，在此过程中我们在图表数据集上对20亿参数的LVLM进行微调以创建ChartJudge。实验表明，多标准提示法揭示了模型鲁棒性的差距，导致包括专门化的LVLM评判器（如LLaVA-Critic）在内的70亿参数模型性能大幅下降。此外，我们发现我们的小型LVLM（ChartJudge）能够将一个数据集的知识转移到另一个数据集中，使其成为一个更加专业化的模型。我们在不同图表类型和查询复杂度上的细致分析提供了关于模型大小、提示设计以及迁移能力之间权衡的实用见解，从而为图表推理任务提供可扩展且低成本的评估方法。我们的代码和数据将会公开发布。
URL
https://arxiv.org/abs/2510.07545
PDF
-
Evaluating Fundus-Specific Foundation Models for Diabetic Macular Edema Detection
2025-10-08 17:41:02Franco Javier Arellano, Jos\'e Ignacio OrlandoAbstract
Diabetic Macular Edema (DME) is a leading cause of vision loss among patients with Diabetic Retinopathy (DR). While deep learning has shown promising results for automatically detecting this condition from fundus images, its application remains challenging due the limited availability of annotated data. Foundation Models (FM) have emerged as an alternative solution. However, it is unclear if they can cope with DME detection in particular. In this paper, we systematically compare different FM and standard transfer learning approaches for this task. Specifically, we compare the two most popular FM for retinal images--RETFound and FLAIR--and an EfficientNet-B0 backbone, across different training regimes and evaluation settings in IDRiD, MESSIDOR-2 and OCT-and-Eye-Fundus-Images (OEFI). Results show that despite their scale, FM do not consistently outperform fine-tuned CNNs in this task. In particular, an EfficientNet-B0 ranked first or second in terms of area under the ROC and precision/recall curves in most evaluation settings, with RETFound only showing promising results in OEFI. FLAIR, on the other hand, demonstrated competitive zero-shot performance, achieving notable AUC-PR scores when prompted appropriately. These findings reveal that FM might not be a good tool for fine-grained ophthalmic tasks such as DME detection even after fine-tuning, suggesting that lightweight CNNs remain strong baselines in data-scarce environments.
Abstract (translated)
糖尿病性黄斑水肿（DME）是糖尿病视网膜病变（DR）患者视力下降的主要原因。尽管深度学习在从眼底图像自动检测这种状况方面显示出有希望的结果，但由于注释数据的可用性有限，其应用仍然面临挑战。基础模型（FM）作为替代解决方案已崭露头角。然而，它们是否能够处理特定的DME检测任务尚不清楚。在这篇论文中，我们系统地比较了不同基础模型和标准迁移学习方法在此任务上的表现。具体来说，我们在IDRiD、MESSIDOR-2 和 OCT-and-Eye-Fundus-Images (OEFI) 数据集的不同训练制度和评估设置下，对两种最流行的眼底图像基础模型——RETFound和FLAIR，以及一个EfficientNet-B0骨干进行了比较。结果显示，尽管规模庞大，基础模型在此任务中并不能始终如一地优于微调的卷积神经网络（CNN）。特别是，在大多数评估设置中，EfficientNet-B0在ROC曲线下的面积（AUC）和精确度/召回率曲线下面积（AUC-PR）方面排名第一或第二，而RETFound仅在OEFI数据集上表现出有前景的结果。另一方面，FLAIR展示了竞争性的零样本性能，在适当提示的情况下取得了显著的AUC-PR分数。这些发现表明，即使经过微调，基础模型可能也不是处理如DME检测这种精细粒度眼科任务的良好工具，这暗示轻量级CNN在数据稀缺环境中仍然是强大的基线模型。
URL
https://arxiv.org/abs/2510.07277
PDF
