Filters








9,211,199 Hits in 6.2 sec

Objects as Points [article]

Xingyi Zhou, Dequan Wang, Philipp Krähenbühl
2019 arXiv   pre-print
We model an object as a single point --- the center point of its bounding box.  ...  Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose.  ...  Conclusion In summary, we present a new representation for objects: as points.  ... 
arXiv:1904.07850v2 fatcat:sqvmgaq35rfx7o4uvmo4d5agle

Tracking Objects as Points [article]

Xingyi Zhou, Vladlen Koltun, Philipp Krähenbühl
2020 arXiv   pre-print
Nowadays, tracking is dominated by pipelines that perform object detection followed by temporal association, also known as tracking-by-detection.  ...  Tracking has traditionally been the art of following interest points through space and time. This changed with the rise of powerful deep networks.  ...  .: Objects as points. arXiv:1904.07850 (2019) 57. Zhu, B., Jiang, Z., Zhou, X., Li, Z., Yu, G.: Class-balanced grouping and sampling for point cloud 3D object detection. arXiv:1908.09492 (2019) 58.  ... 
arXiv:2004.01177v2 fatcat:7nv6ylcf7rfgpfzspdiddp2vzi

Points as Queries: Weakly Semi-supervised Object Detection by Points [article]

Liangyu Chen, Tong Yang, Xiangyu Zhang, Wei Zhang, Jian Sun
2021 arXiv   pre-print
We propose a novel point annotated setting for the weakly semi-supervised object detection task, in which the dataset comprises small fully annotated images and large weakly annotated images by points.  ...  To solve this, we introduce a new detector, Point DETR, which extends DETR by adding a point encoder.  ...  In object detection, we define point annotation as follows: it locates on the object and takes object class as its category.  ... 
arXiv:2104.07434v1 fatcat:g3msreg3hfew7ilh3qidemcxcq

The Renormalization Fixed Point as a Mathematical Object [chapter]

R. P. Langlands
2005 Twenty Years of Bialowieza: A Mathematical Anthology  
regarded as points in the space, but perhaps in a very different part, where the coordinates have quite a different interpretation.  ...  One might wish, as in the early paper [W] of K. G.  ...  It will be defined as an L 2 -space with respect to one of the measures µ S C . The space H introduced, there is a second, more recondite, collection of objects to be defined.  ... 
doi:10.1142/9789812701244_0009 fatcat:snzk3c6hgvd7nkjwwirs4p3dau

Fabric Defect Detection based on Improved Object as Point

Yuan He, Francis Eng Hock Tay, Xin-Yue Huan
2021 Zenodo  
In this paper, we represent a realtime automated detection method based on object as point. This work makes three attributions.  ...  To address the problem, CenterNet method uses bounding box's center point to represent the object, and then on the basis of the point, other attributes of the object such as size and dimension are regressed  ...  And then three different convolutional layers are used to determine the object as point with classification information and its center offset as well as box size.  ... 
doi:10.5281/zenodo.5077222 fatcat:u2gpbiq7sretpd7s6qx3nohrwu

Object Representations as Fixed Points: Training Iterative Refinement Algorithms with Implicit Differentiation [article]

Michael Chang, Thomas L. Griffiths, Sergey Levine
2022 arXiv   pre-print
This property enables the application of such methods to infer representations of sets of entities, such as objects in physical scenes, structurally resembling clustering algorithms in latent space.  ...  Iterative refinement as a fixed point procedure While in general we still lack the theory to understand whether and how the slots of slot attention optimize a well-defined objective like the ELBO, and  ...  Figure 3 : 3 Figure 3: Can slot attention by trained as a fixed point operation with implicit differentiation?  ... 
arXiv:2207.00787v1 fatcat:xq76bojihvcazim5iokadankjy

Fabric Defect Detection based on Improved Object as Point

Yuan He, Xin-Yue Huang, Francis Eng Hock Tay
2021 International Journal of Computer Science & Information Technology (IJCSIT)  
In this paper, we represent a realtime automated detection method based on object as point. This work makes three attributions.  ...  To address the problem, CenterNet method uses bounding box's center point to represent the object, and then on the basis of the point, other attributes of the object such as size and dimension are regressed  ...  And then three different convolutional layers are used to determine the object as point with classification information and its center offset as well as box size.  ... 
doi:10.5121/ijcsit.2021.13301 fatcat:fjwfp2wuqzghvmyg7ajlfo3yja

Predicting Arbitrary-Oriented Objects as Points in Remote Sensing Images

Jian Wang, Le Yang, Fan Li
2021 Remote Sensing  
By directly predicting the center point, orientation, and corresponding height and width of the object, our methods can simplify the design of the model and reduce the computations related to anchors.  ...  In this paper, we convert the detection problem to a center point prediction problem, where the pre-defined anchors can be discarded.  ...  Further, CenterNet [37] models an object as a single point, then regresses the bbox parameters from this point.  ... 
doi:10.3390/rs13183731 fatcat:sjnnb3yxgjfptgsdivgxnjqqqu

The Bloch point in uniaxial ferromagnets as a quantum mechanical object

Andriy Shevchenko, Maksym Barabash
2014 Nanoscale Research Letters  
Quantum effects such as tunneling through pinning barrier of the Bloch Point and over-barrier reflection from the defect potential of one have been investigated in ferromagnets with uniaxial strong magnetic  ...  (BPs; intersection point of two BL parts) [1].  ...  And, as we shall see bellow, the potential of interaction between the BP and a defect has a rather simple form.  ... 
doi:10.1186/1556-276x-9-132 pmid:24646347 pmcid:PMC3994561 fatcat:pi5rei34ozfs3lc2ook3r55fra

Objective Response Rate As a Surrogate End Point: A Commentary

Thomas R. Fleming
2005 Journal of Clinical Oncology  
VOLUME 23 NUMBER 22 AUGUST 1 2005 JOURNAL OF CLINICAL ONCOLOGY Objective Response Rate As a Surrogate End Point: A Commentary Thomas R.  ...  Fleming, University of Washington, Seattle, WA Bruzzi et al’ have evaluated objective response rate (ORR) as a potential surrogate end point for survival in metastatic breast cancer patients.  ... 
doi:10.1200/jco.2005.92.008 pmid:15955898 fatcat:oonitoixzfar7jbanboomvnz4y

Object recognition using point uncertainty regions as pose uncertainty regions

Ilan Shimshoni, Aviva Sasporta
2006 Image and Vision Computing  
From them the pose of the object is estimated and stored in a lookup table. When two similar poses are found the pose is assumed to be correct and the hypothesis is verified.  ...  In this paper, a recognition algorithm based on point features is presented. In this algorithm sets of hypothesized matches between model and image points are generated.  ...  These points do not have to be actual feature points of the object and are not detected in the image, but as they lie within the convex hull of the object, their projection will lie inside the image of  ... 
doi:10.1016/j.imavis.2005.11.003 fatcat:xlic5pobdvg5zmjvwraw7qoeqq

Infants' understanding of the point gesture as an object-directed action

Amanda L. Woodward, Jose J. Guajardo
2002 Cognitive development  
The results of Study 2 suggest that between these ages, infants' own use of object-directed points is related to their understanding of the points of others as object-directed.  ...  These findings suggest that between 9 and 12 months, infants come to understand pointing as an object-directed action.  ...  of pointing as an object-directed action.  ... 
doi:10.1016/s0885-2014(02)00074-6 fatcat:unszqv3ahzdyrkusfl7o3zhqqy

Segment as Points for Efficient Online Multi-Object Tracking and Segmentation [article]

Zhenbo Xu, Wei Zhang, Xiao Tan, Wei Yang, Huan Huang, Shilei Wen, Errui Ding, Liusheng Huang
2020 arXiv   pre-print
Current multi-object tracking and segmentation (MOTS) methods follow the tracking-by-detection paradigm and adopt convolutions for feature extraction.  ...  Furthermore, multiple informative data modalities are converted into point-wise representations to enrich point-wise features.  ...  Visualizing critical points. We visualize critical foreground points as well as critical environment points in Fig. 6 .  ... 
arXiv:2007.01550v1 fatcat:2v76vralhzeslchosnjej2okva

How Objects Categorize the Human Brain: EEG and fMRI as Analysis Point

Rigoberto Cerino, Sergio Vergara
2020 Research on computing science  
An analysis of works that perform semantic classifications of objects such as houses, faces, tools, buildings is presented.  ...  This paper aims to study the state of the art of research based on the reading and processing of brain signals, which allow categorization of different concepts, in multiple aspects, such as images or  ...  In [9] , the neural patterns associated with individual objects as well as with categories of objects were identified using a machine learning algorithm applied to activation distributed throughout the  ... 
dblp:journals/rcs/CerinoV20 fatcat:5djfnq52tjbh3m2buhmr654v6i

Science as a "fixed point"? Quantification and boundary objects in international climate politics

Bård Lahn, Göran Sundqvist
2017 Environmental Science and Policy  
The socalled "Bali Box" gained a prominent role in negotiations leading up to the Copenhagen Conference in 2009, as it was widely seen as providing a "fixed point"a quantified scientific answer to the  ...  The paper employs the notion of boundary objects to analyse the history of the Bali Box, and argues that climate politics will benefit from a richer understanding of the interplay between science and policy  ...  object, are all rendered invisible, while the organization and the object are perceived as representing "pure science".  ... 
doi:10.1016/j.envsci.2016.11.001 fatcat:nca6alvdibad5gfysmkcssmiqu
« Previous Showing results 1 — 15 out of 9,211,199 results