Article
Computer Science, Artificial Intelligence
Jingang Tan, Kangru Wang, Lili Chen, Guanghui Zhang, Jiamao Li, Xiaolin Zhang
Summary: This paper proposes a novel and robust 3D point cloud segmentation framework HCFS3D, which can perform semantic and instance segmentation simultaneously. By using methods like Adaptive Smooth Loss and conditional random fields, the framework shows superior performance in experiments.
IMAGE AND VISION COMPUTING
(2021)
Article
Engineering, Electrical & Electronic
Manonmani Arunkumar, Vijayakumari Pushparaj
Summary: The proposed two-level object segmentation framework utilizes a superpixel-based object boundary gimmicking algorithm and optimized conditional random field algorithm to achieve competitive results in terms of image visualization and performance evaluation.
SIGNAL IMAGE AND VIDEO PROCESSING
(2021)
Article
Environmental Sciences
Shuyang Wang, Xiaodong Mu, Dongfang Yang, Hao He, Peng Zhao
Summary: In this study, an inner convolution integrated encoder-decoder network with directional conditional random fields post-processing was proposed to extract roads from remote sensing images. The approach achieved high-quality road segmentation and connectivity, addressing the problem caused by occlusions.
Article
Computer Science, Artificial Intelligence
Libin Jiao, Lianzhi Huo, Changmiao Hu, Ping Tang
Summary: Refined UNet v3 upgrades the bilateral message-passing kernel and the efficient implementation of Gaussian filtering in the CRF layer, effectively capturing ambiguous edges and accelerating the message-passing procedure. Experimental results demonstrate that the proposed update outperforms its counterpart in terms of detecting vague edges, shadow retrieval, and isolated redundant regions, and it is practically efficient in our TensorFlow implementation.
Article
Computer Science, Artificial Intelligence
Boxiang Zhang, Zunran Wang, Yonggen Ling, Yuanyuan Guan, Shenghao Zhang, Wenhui Li, Lei Wei, Chunxu Zhang
Summary: Transparent object segmentation is a challenging task due to the lack of texture. Shape information plays a critical role in this task. To address this issue, the researchers propose an operation called Patch-wise Weight Shuffle and design a network called ShuffleTrans that performs better in shape recognition. Experimental results on multiple datasets demonstrate the effectiveness of the method in transparent object segmentation.
Article
Ecology
Laura Martinez-Sanchez, Daniele Borio, Raphael d'Andrimont, Marijn van der Velde
Summary: This article investigates the method of estimating tree distances using variations in the skyline of landscape photos. By extracting skyline height and applying various metrics, the study reveals distance-related information.
ECOLOGICAL INFORMATICS
(2022)
Article
Geography, Physical
Shouji Du, Shihong Du, Bo Liu, Xiuyuan Zhang
Summary: This study proposes a semantic segmentation method for VHR images by combining a deep learning semantic segmentation model and object-based image analysis, which aims to capture precise outlines of ground objects and explore context information, achieving competitive overall accuracies for Vaihingen and Potsdam datasets.
INTERNATIONAL JOURNAL OF DIGITAL EARTH
(2021)
Article
Computer Science, Artificial Intelligence
Hongkai Lin, Wentian Xin, Shun Chang, Qianxue Yang, Qiguang Miao, Ruyi Liu, Liang Chang
Summary: This paper proposes a novel network structure, SWHF-Net, to address the issues in semantic segmentation, including underutilization of backbone-derived features and mismatch between small objects and large-scale encodings. SWHF-Net consists of ST-FPM and HF2M modules, which utilize feature transformation and hierarchical fusion to improve the semantic representation of multi-scale objects and enhance computational efficiency.
Article
Computer Science, Information Systems
Hafeez Ur Rehman, Nudrat Nida, Syed Adnan Shah, Wakeel Ahmad, Muhammad Imran Faizi, Syed Muhammad Anwar
Summary: Melanoma, a major cause of death worldwide, is challenging to diagnose and treat. This study proposes a deep learning method that combines image processing, localization, and segmentation to accurately detect and segment melanoma regions.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Jian Ji, Rui Shi, Sitong Li, Peng Chen, Qiguang Miao
Summary: This article proposes a new semantic segmentation method that enhances the model's ability to locate object boundaries by introducing cascaded CRFs into the decoder and fusing the output with the last decoder's output, resulting in more accurate semantic segmentation results.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2021)
Article
Chemistry, Analytical
Chin-Chun Chang, Yen-Po Wang, Shyi-Chyi Cheng
Summary: This paper proposes a method combining Mask R-CNN and PreCNN for fish segmentation in sonar images, improving accuracy and applicability. Using the PreCNN network to extract feature maps, providing standardized inputs for Mask R-CNN, making it better suited for different fish farming environments.
Article
Computer Science, Artificial Intelligence
Tianfei Zhou, Fatih Porikli, David J. Crandall, Luc Van Gool, Wenguan Wang
Summary: Video segmentation is crucial in various practical applications such as enhancing visual effects in movies, understanding scenes in autonomous driving, and creating virtual background in video conferencing. Deep learning-based approaches have shown promising performance in video segmentation. This survey comprehensively reviews two main research lines - generic object segmentation and video semantic segmentation - by introducing their task settings, background concepts, need, development history, and challenges. Representative literature and datasets are also discussed, and the reviewed methods are benchmarked on well-known datasets. Open issues and opportunities for further research are identified, and a public website is provided to track developments in this field: https://github.com/tfzhou/VS-Survey.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Computer Science, Artificial Intelligence
Wei Liu, Ding Li, Hongqi Su
Summary: This study proposed a semantic segmentation framework named hierarchical attention network assembling, which utilizes auxiliary information of different levels corresponding to the hidden and explicit features in the cognitive system, and further processes hidden information to assist the semantic segmentation task.
COGNITIVE COMPUTATION
(2021)
Article
Environmental Sciences
Yingying Kong, Qiupeng Li
Summary: This paper proposes a polarization SAR image semantic segmentation method based on a dual-channel multi-size fully connected convolutional conditional random field. By inputting the full-polarization SAR image and the corresponding optical image simultaneously, integrating multi-size inputs, and introducing the importance of features, the accuracy of image segmentation is improved.
Article
Computer Science, Artificial Intelligence
Zhuo Zheng, Yanfei Zhong, Junjue Wang, Ailong Ma, Liangpei Zhang
Summary: In this paper, a foreground-aware relation network (FarSeg++) is proposed to address the issues of scale variation, large intra-class variance of background, and foreground-background imbalance in high spatial resolution remote sensing imagery. The network improves the discrimination of foreground features, achieves balanced optimization, and enhances objectness representation. Experimental results demonstrate that FarSeg++ outperforms state-of-the-art semantic segmentation methods and achieves a better trade-off between speed and accuracy.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Article
Automation & Control Systems
Pau Rodriguez, Guillem Cucurull, Jordi Gonzalez, Josep M. Gonfaus, Kamal Nasrollahi, Thomas B. Moeslund, F. Xavier Roca
Summary: This paper proposes an automatic system for pain assessment, which outperforms the latest techniques by feeding the raw frames to deep learning models and considering the temporal relation and whole image. The research achieves competitive results in the UNBC-McMaster Shoulder Pain Expression Archive Database and the Cohn Kanade+ facial expression database.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Article
Computer Science, Artificial Intelligence
Meysam Madadi, Sergio Escalera, Xavier Baro, Jordi Gonzalez
Summary: This study introduces a novel hierarchical tree-like structured CNN to address the 3D pose estimation of human hands, training branches to specialize in local poses and fusing features to learn higher order dependencies among joints. Furthermore, a non-rigid data augmentation approach is employed to increase training depth data. Experimental results show competitive performance on various datasets.
IET COMPUTER VISION
(2022)
Article
Computer Science, Artificial Intelligence
Yecong Wan, Yuanshuo Cheng, Mingwen Shao, Jordi Gonzalez
Summary: In this paper, a novel spatially-adaptive network SANet is proposed for simultaneous rain removal and illumination enhancement. A contrastive loss and a new synthetic dataset DarkRain are introduced to boost the development of rain image restoration algorithms.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Information Systems
Vacit Oguz Yazici, Longlong Yu, Arnau Ramisa, Luis Herranz, Joost van de Weijer
Summary: Computer vision has made progress in the online fashion retail industry by proposing a model that utilizes Graph Convolutional Networks (GCN) to detect fashion products in boundary boxes. Compared to the state-of-the-art approach, this method performs better in scenarios where title-input is missing and during cross-dataset evaluation.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Parichehr Behjati, Pau Rodriguez, Carles Fernandez, Isabelle Hupont, Armin Mehri, Jordi Gonzalez
Summary: This study proposes a computationally efficient and accurate single image super-resolution network called DiVANet. By introducing a directional variance attention mechanism and a residual attention feature group, the network is able to improve the performance and efficiency of image recovery.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Artificial Intelligence
Lu Yu, Xialei Liu, Joost van de Weijer
Summary: This paper addresses the problem of catastrophic forgetting in deep neural networks during incremental learning in class-incremental semantic segmentation. A self-training approach is proposed, leveraging unlabeled data for rehearsal of previous knowledge. Experimental results show that maximizing self-entropy and using diverse auxiliary data can significantly improve performance. State-of-the-art results are achieved on Pascal-VOC 2012 and ADE20K datasets.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Information Systems
Wenwen Fu, Zhihong An, Wendong Huang, Haoran Sun, Wenjuan Gong, Jordi Gonzalez
Summary: This study investigates the problem of micro-expression spotting as a frame-by-frame micro-expression classification problem and proposes an effective spotting model. The experimental results demonstrate that the proposed method outperforms the state-of-the-art method in terms of overall F-scores on the CAS(ME)2 and SAMM Long Videos databases.
Article
Computer Science, Artificial Intelligence
Marc Masana, Xialei Liu, Bartlomiej Twardowski, Mikel Menta, Andrew D. Bagdanov, Joost van de Weijer
Summary: For future learning systems, incremental learning is desirable due to its efficient resource usage, reduced memory usage, and resemblance to human learning. The main challenge for incremental learning is catastrophic forgetting. This paper provides a comprehensive survey of existing class-incremental learning methods for image classification and performs extensive experimental evaluations on thirteen methods.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Senmao Li, Joost van de Weijer, Yaxing Wang, Fahad Shahbaz Khan, Meiqin Liu, Jian Yang
Summary: Recent advances in 3D-aware generative models combined with Neural Radiance Fields have achieved impressive results in 3D consistent multi-class image-to-image translation. To address the unrealistic shape/identity change in 2D-I2I translation, the learning process is divided into a multi-class 3D-aware GAN step and a 3D-aware I2I translation step, with novel techniques proposed to reduce view-consistency problems.
2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Aitor Alvarez-Gila, Joost van de Weijer, Yaxing Wang, Estibaliz Garrote
Summary: MVMO is a synthetic dataset with high object density and wide camera baselines, enabling research in multi-view semantic segmentation and cross-view semantic transfer. New research is needed to utilize the information from multi-view setups effectively.
2022 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Vacit Oguz Yazici, Joost Van De Weijer, Longlong Yu
Summary: This paper investigates the problem of multi-label image classification and proposes an enhanced transformer model that utilizes primal object queries to improve model performance and convergence speed.
2022 26TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR)
(2022)
Proceedings Paper
Computer Science, Theory & Methods
Hector Laria, Yaxing Wang, Joost van de Weijer, Bogdan Raducanu
Summary: GANs have matured in recent years and can generate high-resolution, realistic images. This paper focuses on transferring from high-quality pretrained unconditional GANs to conditional GANs, proposing hyper-modulated generative networks for architectural adaptation and introducing self-initialization and contrastive loss for improved transfer efficiency.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW 2022
(2022)
Proceedings Paper
Computer Science, Theory & Methods
Bojana Gajic, Ariel Amato, Ramon Baldrich, Joost van de Weijer, Carlo Gatta
Summary: Most popular metric learning losses are not directly related to the evaluation metrics used to assess their performance. However, training a metric learning model by maximizing the area under the ROC curve can induce a suitable implicit ranking for retrieval problems. By proposing an approximated and derivable AUC loss, state-of-the-art performance is achieved on large scale retrieval benchmark datasets.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW 2022
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Javad Zolfaghari Bengar, Joost van de Weijer, Laura Lopez Fuentes, Bogdan Raducanu
Summary: In real-world scenarios, imbalanced class distribution in datasets further complicates the active learning process. To address this issue, we propose an optimization framework considering class-balancing, which can effectively improve the performance of active learning methods.
2022 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2022)
(2022)
Article
Computer Science, Information Systems
Parichehr Behjati, Pau Rodriguez, Carles Fernandez Tena, Armin Mehri, F. Xavier Roca, Seiichi Ozawa, Jordi Gonzalez
Summary: This study focuses on single image super-resolution based on deep convolutional neural networks (CNNs), proposing a novel Frequency-based Enhancement Block (FEB) to enhance high-frequency information and recover finer details. Experimental results show that replacing commonly used SR blocks with FEB improves reconstruction error and reduces the number of parameters in the model.