Filters








73 Hits in 1.4 sec

OCNet: Object Context Network for Scene Parsing [article]

Yuhui Yuan, Lang Huang, Jianyuan Guo, Chao Zhang, Xilin Chen, Jingdong Wang
2021 arXiv   pre-print
Motivated by the fact that the category of each pixel is inherited from the object it belongs to, we define the object context for each pixel as the set of pixels that belong to the same category as the  ...  In this paper, we address the semantic segmentation task with a new context aggregation scheme named object context, which focuses on enhancing the role of object information.  ...  PASCAL-Context (Mottaghi et al., 2014 ) is a challenging scene parsing dataset that contains (a) OCNet BackBone Object Context Module Classifier (b) Base-OC 3x3 Conv OCP (c) Pyramid-OC 1x1  ... 
arXiv:1809.00916v4 fatcat:h6r25be3hbbbbkcipf4pxea56i

Deep Grouping Model for Unified Perceptual Parsing [article]

Zhiheng Li, Wenxuan Bao, Jiayang Zheng, Chenliang Xu
2020 arXiv   pre-print
When evaluating the model on the recent Broden+ dataset for the unified perceptual parsing task, it achieves state-of-the-art results while having a small computational overhead compared to other contextual-based  ...  Examples can be found in the classical hierarchical superpixel segmentation or image parsing works.  ...  Similarly, global scene context can also be propagated down to lower-level graph containing objects.  ... 
arXiv:2003.11647v1 fatcat:bopfoo3hcreztbxmgbjfdktwli

Deep Grouping Model for Unified Perceptual Parsing

Zhiheng Li, Wenxuan Bao, Jiayang Zheng, Chenliang Xu
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
When evaluating the model on the recent Broden+ dataset for the unified perceptual parsing task, it achieves state-ofthe-art results while having a small computational overhead compared to other contextual-based  ...  Examples can be found in the classical hierarchical superpixel segmentation or image parsing works.  ...  Similarly, global scene context can also be propagated down to lower-level graph containing objects.  ... 
doi:10.1109/cvpr42600.2020.00411 dblp:conf/cvpr/LiBZX20 fatcat:mmcaro5id5axvpquq3auawhyja

Gated Path Selection Network for Semantic Segmentation [article]

Qichuan Geng, Hong Zhang, Xiaojuan Qi, Ruigang Yang, Zhong Zhou, Gao Huang
2020 arXiv   pre-print
In this paper, we develop a novel network named Gated Path Selection Network (GPSNet), which aims to learn adaptive receptive fields.  ...  The derived adaptive receptive fields are data-dependent, and are flexible that can model different object geometric transformations.  ...  The scene parsing dataset ADE20K contains 150 classes and diverse complex scenes with 1,038 imagelevel categories. It needs to parse both objects and stuff.  ... 
arXiv:2001.06819v1 fatcat:y6w5amzi4vcyxpa36ttsabbtvq

Deep Multiphase Level Set for Scene Parsing [article]

Pingping Zhang and Wei Liu and Yinjie Lei and Hongyu Wang and Huchuan Lu
2019 arXiv   pre-print
Recently, Fully Convolutional Network (FCN) seems to be the go-to architecture for image segmentation, including semantic scene parsing.  ...  However, it is difficult for a generic FCN to discriminate pixels around the object boundaries, thus FCN based methods may output parsing results with inaccurate boundaries.  ...  [32] propose the pyramid object context to model the category dependencies. Fu et al.  ... 
arXiv:1910.03166v2 fatcat:7viff6ta75hy7fgg5rnxcya2wa

CaseNet: Content-Adaptive Scale Interaction Networks for Scene Parsing [article]

Xin Jin, Cuiling Lan, Wenjun Zeng, Zhizheng Zhang, Zhibo Chen
2020 arXiv   pre-print
In this paper, we propose a Content-Adaptive Scale Interaction Network (CASINet) to exploit the multi-scale features for scene parsing.  ...  Objects at different spatial positions in an image exhibit different scales. Adaptive receptive fields are expected to capture suitable ranges of context for accurate pixel level semantic prediction.  ...  Scene parsing has achieved great progress with the development of Fully Convolutional Networks (FCNs) [11] .  ... 
arXiv:1904.08170v3 fatcat:rxeulrn225gerdd5mf5mgamqum

Deep Learning Model with GA based Feature Selection and Context Integration [article]

Ranju Mandal, Basim Azam, Brijesh Verma, Mengjie Zhang
2022 arXiv   pre-print
Stanford Background and CamVid benchmark image parsing datasets were used for our model evaluation, and our model shows promising results.  ...  Since its inception, Many top-performing methods for image segmentation are based on deep CNN models.  ...  Predictive scene parsing models: Spatio-Temporally Coupled Generative Adversarial Networks (STC-GAN) [20] combines a future frame generation model with a predictive scene parsing model.  ... 
arXiv:2204.06189v1 fatcat:qkyclvhyrzed7lt2cimvylj2ou

Deep Dual-resolution Networks for Real-time and Accurate Semantic Segmentation of Road Scenes [article]

Yuanduo Hong, Huihui Pan, Weichao Sun, Yisong Jia
2021 arXiv   pre-print
Semantic segmentation is a key technology for autonomous vehicles to understand the surrounding scenes.  ...  Using light-weight architectures (encoder-decoder or two-pathway) or reasoning on low-resolution images, recent methods realize very fast scene parsing, even running at more than 100 FPS on a single 1080Ti  ...  Object Context Network (OCNet) [37] utilizes self-attention mechanism to explore object context which is defined as a set of pixels belonging to the same object category.  ... 
arXiv:2101.06085v2 fatcat:lzmgglakurb4tkvxlwfkfwx33e

ISNet: Integrate Image-Level and Semantic-Level Context for Semantic Segmentation [article]

Zhenchao Jin, Bin Liu, Qi Chu, Nenghai Yu
2021 arXiv   pre-print
First, an image-level context module is designed to capture the contextual information for each pixel in the whole image.  ...  Co-occurrent visual pattern makes aggregating contextual information a common paradigm to enhance the pixel representation for semantic image segmentation.  ...  Since there are only 9K images in the train set of COCOStuff where these images contain 182 semantic classes, COCOStuff is a very challenging benchmark for scene parsing.  ... 
arXiv:2108.12382v1 fatcat:rvkmcmn5t5apbmnezt2k7uzc74

Self-Correction for Human Parsing [article]

Peike Li, Yunqiu Xu, Yunchao Wei, Yi Yang
2019 arXiv   pre-print
Labeling pixel-level masks for fine-grained semantic segmentation tasks, e.g. human parsing, remains a challenging task.  ...  To tackle the problem of learning with label noises, this work introduces a purification strategy, called Self-Correction for Human Parsing (SCHP), to progressively promote the reliability of the supervised  ...  Acknowledgement We thank Ting Liu and Tao Ruan for providing insights and expertise to improve this work.  ... 
arXiv:1910.09777v1 fatcat:oyyexar3ijbmbnmxhnoilfkope

Trans4Trans: Efficient Transformer for Transparent Object and Semantic Scene Segmentation in Real-World Navigation Assistance [article]

Jiaming Zhang, Kailun Yang, Angela Constantinescu, Kunyu Peng, Karin Müller, Rainer Stiefelhagen
2021 arXiv   pre-print
In this paper, we build a wearable system with a novel dual-head Transformer for Transparency (Trans4Trans) perception model, which can segment general- and transparent objects.  ...  Meanwhile, the Tran4Trans model has outstanding performances on driving scene datasets.  ...  Our proposed approach outperforms both these aforementioned competitive networks such as OCNet and transformer-based encoder-decoder architectures such as PVT.  ... 
arXiv:2108.09174v1 fatcat:raco7sb4ybcerbqbpppixxn67q

AttaNet: Attention-Augmented Network for Fast and Accurate Scene Parsing [article]

Qi Song and Kangfu Mei and Rui Huang
2021 arXiv   pre-print
In this paper, we propose a new model, called Attention-Augmented Network (AttaNet), to capture both global context and multilevel semantics while keeping the efficiency high.  ...  Two factors have proven to be very important to the performance of semantic segmentation models: global context and multi-level semantics.  ...  Acknowledgments This work is supported in part by funding from Shenzhen Institute of Artificial Intelligence and Robotics for Society, and Shenzhen NSF JCYJ20190813170601651.  ... 
arXiv:2103.05930v1 fatcat:icpiordlujbz3hkfhjhprp7kdy

Evaluation of Deep Learning Segmentation Models for Detection of Pine Wilt Disease in Unmanned Aerial Vehicle Images

Lang Xia, Ruirui Zhang, Liping Chen, Longlong Li, Tongchuan Yi, Yao Wen, Chenchen Ding, Chunchun Xie
2021 Remote Sensing  
., fully convolutional networks for semantic segmentation, DeepLabv3+, and PSPNet) were trained and evaluated.  ...  Also, an atrous spatial pyramid pooling module encoded multiscale context information, and the encoder–decoder architecture recovered location/spatial information, being the best architecture for segmenting  ...  Acknowledgments: We thank the Landscape and Forestry Bureau of Qindao for supporting the collection of UAV data. Conflicts of Interest: The authors declare no conflict of interest.  ... 
doi:10.3390/rs13183594 fatcat:dykrbm4xezdorpai3vngnexkmq

Dual Attention Network for Scene Segmentation [article]

Jun Fu, Jing Liu, Haijie Tian, Yong Li, Yongjun Bao, Zhiwei Fang, Hanqing Lu
2019 arXiv   pre-print
Unlike previous works that capture contexts by multi-scale features fusion, we propose a Dual Attention Networks (DANet) to adaptively integrate local features with their global dependencies.  ...  We achieve new state-of-the-art segmentation performance on three challenging scene segmentation datasets, i.e., Cityscapes, PASCAL Context and COCO Stuff dataset.  ...  The concurrent work OCNet [27] adopts self-attention mechanism with ASPP to exploit context dependencies.  ... 
arXiv:1809.02983v4 fatcat:6jowndquhbg3tgcuohmtuwazmu

Global-and-Local Context Network for Semantic Segmentation of Street View Images

Chih-Yang Lin, Yi-Cheng Chiu, Hui-Fuang Ng, Timothy K. Shih, Kuan-Hung Lin
2020 Sensors  
objects in a scene, thus reducing segmentation errors.  ...  Semantic segmentation of street view images is an important step in scene understanding for autonomous vehicle systems.  ...  Self-attention is also used in OCNet [23] to learn pixel-level object context information to enhance context aggregation.  ... 
doi:10.3390/s20102907 pmid:32455537 fatcat:tmlnyjqzivbibeqklgyot3bxdm
« Previous Showing results 1 — 15 out of 73 results