A framework merging different knowledge types to improve model performance.
Yaomin Huang, Zaomin Yan, Chaomin Shen
― 5 min read
Computer Science - Computer Vision and Pattern Recognition
RSSA new method improves gaze target detection with less labeled data.
Francesco Tonini, Nicola Dall'Asen, Lorenzo Vaquero
― 6 min read
A new approach enhances deep learning model performance amidst noise.
Seyedarmin Azizi, Mohammad Erfan Sadeghi, Mehdi Kamal
― 5 min read
A new framework improves pixel labeling by addressing uncertainty in semantic segmentation.
Xiaoke Hao, Shiyu Liu, Chuanbo Feng
― 6 min read
This study assesses the effectiveness of pre-trained models in Earth Observation applications.
Jose Sosa, Mohamed Aloulou, Danila Rukhovich
― 6 min read
Temporal2Seq framework streamlines multiple video understanding tasks into one model.
Min Yang, Zichen Zhang, Limin Wang
― 8 min read
TAKFL optimizes knowledge sharing in federated learning for diverse device capabilities.
Mahdi Morafah, Vyacheslav Kungurtsev, Hojin Chang
― 6 min read
A method that aligns 3D shapes with 2D images without matched points.
Jingwei Song, Maani Ghaffari
― 6 min read
Explore the essential concepts of molecular physics and their practical applications.
Jun Liu, Geng Yuan, Weihao Zeng
― 4 min read
This new method streamlines image generation in AI models, improving efficiency and speed.
Seongmin Hong, Suh Yoon Jeon, Kyeonghyun Lee
― 6 min read
A new framework enhances video-language dataset quality through iterative refinement.
Xiao Wang, Jianlong Wu, Zijia Lin
― 5 min read
Combining street view images with data to analyze building exteriors.
Zongrong Li, Yunlei Su, Chenyuan Zhu
― 6 min read
A model to assess segmentation quality without ground truth benchmarks.
Ahjol Senbi, Tianyu Huang, Fei Lyu
― 8 min read
MedCLIP-SAMv2 improves tumor detection using advanced segmentation techniques and minimal labeled data.
Taha Koleilat, Hojat Asgariandehkordi, Hassan Rivaz
― 5 min read
A look into how CNNs learn image features and their universal similarities.
Florentin Guth, Brice Ménard
― 7 min read
Researchers use CRISP to improve biodiversity tracking through better image analysis.
Andy V. Huynh, Lauren E. Gillespie, Jael Lopez-Saucedo
― 6 min read
A new index helps assess diversity in AI-generated medical images.
Mohammed Talha Alam, Raza Imam, Mohammad Areeb Qazi
― 9 min read
New methods speed up video encoding and decoding.
Hao Chen, Saining Xie, Ser-Nam Lim
― 5 min read
A new framework enhances the connection between images and text.
Mayug Maniparambil, Raiymbek Akshulakov, Yasser Abdelaziz Dahou Djilali
― 7 min read
Learn how machine learning models can improve when facing new and unseen data.
Zongbo Han, Jialong Yang, Junfan Li
― 7 min read
A look at the role and methods of diffusion models in image creation.
Zheyuan Zhan, Defang Chen, Jian-Ping Mei
― 7 min read
Exploring methods to improve multimodal models in breaking down visual questions.
Haowei Zhang, Jianzhe Liu, Zhen Han
― 6 min read
A new model generates reports from 3D CT scans efficiently and accurately.
Hao Chen, Wei Zhao, Yingli Li
― 8 min read
A new pipeline for generating 3D models from 2D images efficiently.
Potito Aghilar, Vito Walter Anelli, Michelantonio Trizio
― 5 min read
TrojVLM exposes vulnerabilities in Vision Language Models to backdoor attacks.
Weimin Lyu, Lu Pang, Tengfei Ma
― 7 min read
This study reveals effective methods for recognizing hand gestures through ultrasound imaging.
Keshav Bimbraw, Ankit Talele, Haichong K. Zhang
― 5 min read
A new framework improves data generation across multiple sources using energy-based models.
Shiyu Yuan, Jiali Cui, Hanao Li
― 5 min read
SATA improves the robustness and efficiency of Vision Transformers for image classification tasks.
Nick Nikzad, Yi Liao, Yongsheng Gao
― 4 min read
A new method improves object recognition using masks without detailed labels.
Heeseong Shin, Chaehyun Kim, Sunghwan Hong
― 5 min read
A new method simplifies the removal of unwanted content in visual datasets.
Saehyung Lee, Jisoo Mok, Sangha Park
― 6 min read
Exploring Federated Learning's role in enhancing medical imaging while protecting patient privacy.
Nikolas Koutsoubis, Asim Waqas, Yasin Yilmaz
― 5 min read
A new method helps robots learn tasks using online human videos, reducing training needs.
Homanga Bharadhwaj, Debidatta Dwibedi, Abhinav Gupta
― 6 min read
PPLNs enhance event camera data processing for improved machine vision capabilities.
Chen Song, Zhenxiao Liang, Bo Sun
― 6 min read
A novel approach improves the detection of genuine signatures against forgeries.
Hansong Zhang, Jiangjian Guo, Kun Li
― 5 min read
Analyzing the effects of pruning methods on GoogLeNet's performance and interpretability.
Jonathan von Rad, Florian Seuffert
― 5 min read
Innovative methods for enhancing depth maps vital for augmented and virtual reality.
Marcos V. Conde, Florin-Alexandru Vasluianu, Jinhui Xiong
― 6 min read
FAST improves disease classification using whole slide images with minimal expert input.
Kexue Fu, Xiaoyuan Luo, Linhao Qu
― 5 min read
A method to enhance model performance despite incorrect data labels.
Tong Wei, Hao-Tian Li, Chun-Shu Li
― 7 min read
MedViLaM integrates multiple medical data types for improved analysis and decision-making.
Lijian Xu, Hao Sun, Ziyu Ni
― 5 min read
A new method to speed up diffusion model output without losing quality.
Zhenyu Zhou, Defang Chen, Can Wang
― 7 min read