作者: Bridle 時間: 2025-3-22 00:18
Dealing with Ambiguity in Robotic Grasping via Multiple Predictionsel that predicts a set of grasp hypotheses in under 60?ms, which is critical for real-time robotic applications. The grasp detection accuracy reaches over . for unseen objects, outperforming the current state of the art on this task.作者: 滔滔不絕地講 時間: 2025-3-22 00:36 作者: 勛章 時間: 2025-3-22 07:45
Robust Deep Multi-modal Learning Based on Gated Information Fusion Networkmodality according to the input feature maps to be fused. The combining weights are determined by applying the convolutional layers followed by the sigmoid function to the concatenated intermediate feature maps. The whole network including the CNN backbone and GIF is trained in an end-to-end fashion作者: ironic 時間: 2025-3-22 11:56
Hardware-Aware Softmax Approximation for Deep Neural Networksating cost-intensive operations in Softmax (. exponential and division) with cost-effective operations (. addition and bit shifts). We designed and synthesized a hardware unit for our approximation approach, to estimate the area and energy consumption. In addition, we introduce a training method to 作者: 遠(yuǎn)地點 時間: 2025-3-22 13:59
Video Object Segmentation with Language Referring Expressions. We show that our language-supervised approach performs on par with the methods which have access to a pixel-level mask of the target object on . and is competitive to methods using scribbles on the challenging . dataset.作者: 遠(yuǎn)地點 時間: 2025-3-22 21:07
Nonlinear Subspace Feature Enhancement for Image Set Classificatione of subspace-based classifiers such as sparse representation-based classification. We describe how the structured loss function of NSFE can be optimized in a batch-by-batch fashion by a two-step alternating algorithm. The algorithm makes very few assumptions about the form of the embedding to be le作者: ingrate 時間: 2025-3-22 23:23
Adversarial Learning for Visual Storytelling with Sense Group Partitionup as the unit, we propose to do the paragraph generation at sense group level instead of sentence level. Experiments on the widely-used dataset show that our approach generates higher-quality descriptions than previous baselines.作者: Conflagration 時間: 2025-3-23 01:37 作者: Hemiplegia 時間: 2025-3-23 06:09 作者: 附錄 時間: 2025-3-23 12:08
Learning from PhotoShop Operation Videos: The PSOV Datasetorresponding evaluation metrics. To demonstrate that the PSOV dataset has sufficient data and labeling for data-driven methods, we develop a deep learning based algorithm for the command classification task. We also carry out experiments and analysis with the proposed method to encourage better unde作者: 壁畫 時間: 2025-3-23 14:11 作者: 天然熱噴泉 時間: 2025-3-23 18:22
Symmetry-Aware Face Completion with Generative Adversarial Networksur method is capable of synthesizing semantically valid and visually plausible contents for the missing facial key parts from random mask. In addition, our model outperforms other methods for detail completion of facial components.作者: 高貴領(lǐng)導(dǎo) 時間: 2025-3-23 22:48 作者: babble 時間: 2025-3-24 03:37
Fusing Solar and Stellar Cosmologies,el that predicts a set of grasp hypotheses in under 60?ms, which is critical for real-time robotic applications. The grasp detection accuracy reaches over . for unseen objects, outperforming the current state of the art on this task.作者: 撫育 時間: 2025-3-24 07:43
The Wider Copernican Revolution,rmation and handle the domain distribution mismatch simultaneously. Our key novelties are: (1) a global visual-depth metric construction algorithm that can effectively align RGB and depth data structure; (2) adaptive transformed component extraction for target domain that conditioned on invariant tr作者: Outmoded 時間: 2025-3-24 12:49 作者: 臭名昭著 時間: 2025-3-24 17:55
https://doi.org/10.1007/978-981-13-8699-2ating cost-intensive operations in Softmax (. exponential and division) with cost-effective operations (. addition and bit shifts). We designed and synthesized a hardware unit for our approximation approach, to estimate the area and energy consumption. In addition, we introduce a training method to 作者: obnoxious 時間: 2025-3-24 22:38 作者: opprobrious 時間: 2025-3-25 00:17
Farmers, Fossils, and Heat, July 18–22e of subspace-based classifiers such as sparse representation-based classification. We describe how the structured loss function of NSFE can be optimized in a batch-by-batch fashion by a two-step alternating algorithm. The algorithm makes very few assumptions about the form of the embedding to be le作者: pus840 時間: 2025-3-25 03:22
Harold Rogers Wanless,Emmett Evanoffup as the unit, we propose to do the paragraph generation at sense group level instead of sentence level. Experiments on the widely-used dataset show that our approach generates higher-quality descriptions than previous baselines.作者: 高原 時間: 2025-3-25 09:41 作者: Antigen 時間: 2025-3-25 13:29 作者: 酷熱 時間: 2025-3-25 17:13 作者: Amnesty 時間: 2025-3-25 21:18
https://doi.org/10.1007/978-94-010-2456-3 show that the numerical stability of our algorithm is comparable to the state-of-the-art Perspective-3-Point (.) algorithm?[.], and outperforms the state-of-the-art algorithms of the other three cases. The numerical stability of our algorithm can be further improved by a rough estimation of the rot作者: configuration 時間: 2025-3-26 03:37
The Diaspora and Jewish Characterur method is capable of synthesizing semantically valid and visually plausible contents for the missing facial key parts from random mask. In addition, our model outperforms other methods for detail completion of facial components.作者: HUMID 時間: 2025-3-26 05:52
The Narrow Copernican Revolution,task to be solved, they can have long computation times for more difficult problem instances. Here, we propose two variants of a heuristic solver (primal feasible heuristic), which greedily generate solutions within a bounded amount of time. Evaluations on image and mesh segmentation benchmarks show the high quality of these solutions.作者: Eulogy 時間: 2025-3-26 10:30 作者: Addictive 時間: 2025-3-26 13:22 作者: 固執(zhí)點好 時間: 2025-3-26 18:35 作者: 試驗 時間: 2025-3-26 23:51
Conference proceedings 2019a, in December 2018. The total of 274 contributions was carefully reviewed and selected from 979 submissions during two rounds of reviewing and improvement. The papers focus on motion and tracking, segmentation and grouping, image-based modeling, dep learning, object recognition object recognition, 作者: GRUEL 時間: 2025-3-27 02:21 作者: lethargy 時間: 2025-3-27 08:22
The Diaspora: Origin and Meaninghitecture which can generate more accurate segmentation prediction without heavy calculation burden are adopted. Our experiments on the PASCAL VOC and the MSCOCO datasets demonstrate the benefits of our approach, which accelerate the instance segmentation process with competitive result compared to MaskRCNN. Code is public available (.).作者: rectocele 時間: 2025-3-27 13:08
Continual Occlusion and Optical Flow Estimationme flow serves as an input to occlusion estimation and as a prior in occluded regions, i.e. those without visual correspondences. By continually using the previous frame flow, ContinualFlow performance improves further by 18% on KITTI and 7% on Sintel, achieving top performance on KITTI and Sintel.作者: 驚奇 時間: 2025-3-27 17:14
Fast Single Shot Instance Segmentationhitecture which can generate more accurate segmentation prediction without heavy calculation burden are adopted. Our experiments on the PASCAL VOC and the MSCOCO datasets demonstrate the benefits of our approach, which accelerate the instance segmentation process with competitive result compared to MaskRCNN. Code is public available (.).作者: 獨行者 時間: 2025-3-27 21:21 作者: moribund 時間: 2025-3-28 01:40
Integrating Hellenic and Indian Traditions,ty. We provide experiments that show how our proposed model provides favorable quantitative results compared with the state-of-the-art while simultaneously providing a representation that resides on a low-dimensional interpretable manifold.作者: 休閑 時間: 2025-3-28 03:24
Maria Carolina: Sovereign and Mother,e caricature recognition task. We have verified that both local and global features are crucial for caricature recognition. Moreover, experimental results show that, compared with the state-of-the-art methods, LGDML can obtain superior performance in terms of Rank-1 and Rank-10.作者: 新奇 時間: 2025-3-28 06:56
Gaussian Process Deep Belief Networks: A Smooth Generative Model of Shape with Uncertainty Propagatity. We provide experiments that show how our proposed model provides favorable quantitative results compared with the state-of-the-art while simultaneously providing a representation that resides on a low-dimensional interpretable manifold.作者: Narrative 時間: 2025-3-28 13:59
A Joint Local and Global Deep Metric Learning Method for Caricature Recognitione caricature recognition task. We have verified that both local and global features are crucial for caricature recognition. Moreover, experimental results show that, compared with the state-of-the-art methods, LGDML can obtain superior performance in terms of Rank-1 and Rank-10.作者: gustation 時間: 2025-3-28 16:58 作者: amphibian 時間: 2025-3-28 21:57
Gated Hierarchical Attention for Image Captioningring and neural machine translation, and significantly improves the performance. In this paper, we propose a bottom-up gated hierarchical attention (GHA) mechanism for image captioning. Our proposed model employs a CNN as the decoder which is able to learn different concepts at different layers, and作者: Lobotomy 時間: 2025-3-29 00:26
Dealing with Ambiguity in Robotic Grasping via Multiple Predictionsjects. However, the lack of such intuition in robots makes robotic grasping an exceptionally challenging task. There are often several equally viable options of grasping an object. However, this ambiguity is not modeled in conventional systems that estimate a single, optimal grasp position. We propo作者: OASIS 時間: 2025-3-29 05:13 作者: Left-Atrium 時間: 2025-3-29 07:19
Solving Minimum Cost Lifted Multicut Problems by Node Agglomerationosition or multiple object tracking. Its solutions are decompositions of a graph into an optimal number of segments which are optimized w.r.t. a cost function defined on a superset of the edge set. While the currently available solvers for this problem provide high quality solutions in terms of the 作者: 使饑餓 時間: 2025-3-29 14:06 作者: 側(cè)面左右 時間: 2025-3-29 19:00 作者: 同位素 時間: 2025-3-29 21:15 作者: Anterior 時間: 2025-3-30 03:34 作者: Moderate 時間: 2025-3-30 07:58
Continual Occlusion and Optical Flow Estimationames. The proposed ContinualFlow method estimates occlusions before flow, avoiding the use of flow corrupted by occlusions for their estimation. We show that providing occlusion masks as an additional input to flow estimation improves the standard performance metric by more than 25% on both KITTI an作者: 黃瓜 時間: 2025-3-30 08:31
Adversarial Learning for Visual Storytelling with Sense Group Partitionision and language research, the techniques for sequential vision-to-language are still far away from being perfect. Due to the limitation of maximum likelihood estimation on training, the majority of existing models encourage high resemblance to texts in the training database, which makes the descr作者: aspect 時間: 2025-3-30 14:09
Laser Scar Detection in Fundus Images Using Convolutional Neural Networksbehind circular or irregular scars in the retina. Laser scar detection in fundus images is thus important for automated DR screening. Despite its importance, the problem is understudied in terms of both datasets and methods. This paper makes the first attempt to detect laser-scar images by deep lear作者: 陳腐思想 時間: 2025-3-30 20:07 作者: 傳染 時間: 2025-3-31 00:39 作者: 精密 時間: 2025-3-31 04:33
A Joint Local and Global Deep Metric Learning Method for Caricature Recognitioneen photographs and caricatures, making it nontrivial to match the features of photographs and caricatures. To address the problem, a joint local and global metric learning method (LGDML) is proposed. First, joint local and global feature representation is learnt with convolutional neural networks t作者: wreathe 時間: 2025-3-31 05:08
Fast Single Shot Instance Segmentationg every individual instance (instance segmentation) in a flexible and fast way. In the pipeline of FSSI, the instance segmentation task is divided into three parallel sub-tasks: object detection, semantic segmentation, and direction prediction. The instance segmentation result is then generated from作者: absorbed 時間: 2025-3-31 10:20 作者: 改進(jìn) 時間: 2025-3-31 17:00
Symmetry-Aware Face Completion with Generative Adversarial Networksthout taking these characteristics into account, existing face completion techniques usually fail to produce a photo-realistic result, especially for the missing key components (e.g., eyes and mouths). In this paper, we propose a symmetry-aware face completion method based on facial structural featu作者: fringe 時間: 2025-3-31 17:35 作者: NIL 時間: 2025-4-1 00:07
https://doi.org/10.1007/978-3-030-20870-7artificial intelligence; computer vision; databases; image coding; image processing; image reconstruction作者: sed-rate 時間: 2025-4-1 05:44 作者: Emmenagogue 時間: 2025-4-1 08:07
Computer Vision – ACCV 2018978-3-030-20870-7Series ISSN 0302-9743 Series E-ISSN 1611-3349 作者: 名次后綴 時間: 2025-4-1 12:23
Integrating Hellenic and Indian Traditions,arance, it is possible to generalize to a large range of objects from only small amounts of data. However, shapes represented as silhouette images are challenging to model due to complicated likelihood functions leading to intractable posteriors. In this paper we present a generative model of shapes作者: Dorsal-Kyphosis 時間: 2025-4-1 16:01
