Monocular Depth Estimation (MDE) is a fundamental problem in computer vision with numerous applications. Recently, LIDAR-supervised methods have achieved remarkable per-pixel depth accuracy in outdoor scenes. However, significant errors are typically found in the proximity of depth discontinuities, i.e., depth edges, which often hinder the performance of depth-dependent applications that are sensitive to such inaccuracies, e.g., novel view synthesis and augmented reality. Since direct supervision for the location of depth edges is typically unavailable in sparse LIDAR-based scenes, encouraging the MDE model to produce correct depth edges is not straightforward. In this work we propose to learn to detect the location of depth edges from densely-supervised synthetic data, and use it to generate supervision for the depth edges in the MDE training. %Despite the 'domain gap' between synthetic and real data, we show that depth edges that are estimated directly are significantly more accurate than the ones that emerge indirectly from the MDE training. To quantitatively evaluate our approach, and due to the lack of depth edges ground truth in LIDAR-based scenes, we manually annotated subsets of the KITTI and the DDAD datasets with depth edges ground truth. We demonstrate significant gains in the accuracy of the depth edges with comparable per-pixel depth accuracy on several challenging datasets.
translated by 谷歌翻译
来自单个运动模糊图像的视频重建是一个具有挑战性的问题,可以增强现有的相机的能力。最近,几种作品使用传统的成像和深度学习解决了这项任务。然而,由于方向模糊和噪声灵敏度,这种纯粹 - 数字方法本质上是有限的。一些作品提出使用非传统图像传感器解决这些限制,然而,这种传感器非常罕见和昂贵。为了使这些限制具有更简单的方法,我们提出了一种用于视频重建的混合光学 - 数字方法,其仅需要对现有光学系统的简单修改。在图像采集期间,在镜头孔径中使用学习的动态相位编码以对运动轨迹进行编码,该运动轨迹用作视频重建过程的先前信息。使用图像到视频卷积神经网络,所提出的计算相机以各种编码运动模糊图像的各种帧速率产生锐帧帧突发。与现有方法相比,我们使用模拟和现实世界的相机原型表现了优势和改进的性能。
translated by 谷歌翻译
A core process in human cognition is analogical mapping: the ability to identify a similar relational structure between different situations. We introduce a novel task, Visual Analogies of Situation Recognition, adapting the classical word-analogy task into the visual domain. Given a triplet of images, the task is to select an image candidate B' that completes the analogy (A to A' is like B to what?). Unlike previous work on visual analogy that focused on simple image transformations, we tackle complex analogies requiring understanding of scenes. We leverage situation recognition annotations and the CLIP model to generate a large set of 500k candidate analogies. Crowdsourced annotations for a sample of the data indicate that humans agree with the dataset label ~80% of the time (chance level 25%). Furthermore, we use human annotations to create a gold-standard dataset of 3,820 validated analogies. Our experiments demonstrate that state-of-the-art models do well when distractors are chosen randomly (~86%), but struggle with carefully chosen distractors (~53%, compared to 90% human accuracy). We hope our dataset will encourage the development of new analogy-making models. Website: https://vasr-dataset.github.io/
translated by 谷歌翻译
We introduce MuJoCo MPC (MJPC), an open-source, interactive application and software framework for real-time predictive control, based on MuJoCo physics. MJPC allows the user to easily author and solve complex robotics tasks, and currently supports three shooting-based planners: derivative-based iLQG and Gradient Descent, and a simple derivative-free method we call Predictive Sampling. Predictive Sampling was designed as an elementary baseline, mostly for its pedagogical value, but turned out to be surprisingly competitive with the more established algorithms. This work does not present algorithmic advances, and instead, prioritises performant algorithms, simple code, and accessibility of model-based methods via intuitive and interactive software. MJPC is available at: github.com/deepmind/mujoco_mpc, a video summary can be viewed at: dpmd.ai/mjpc.
translated by 谷歌翻译
Robotic planning in real-world scenarios typically requires joint optimization of logic and continuous variables. A core challenge to combine the strengths of logic planners and continuous solvers is the design of an efficient interface that informs the logical search about continuous infeasibilities. In this paper we present a novel iterative algorithm that connects logic planning with nonlinear optimization through a bidirectional interface, achieved by the detection of minimal subsets of nonlinear constraints that are infeasible. The algorithm continuously builds a database of graphs that represent (in)feasible subsets of continuous variables and constraints, and encodes this knowledge in the logical description. As a foundation for this algorithm, we introduce Planning with Nonlinear Transition Constraints (PNTC), a novel planning formulation that clarifies the exact assumptions our algorithm requires and can be applied to model Task and Motion Planning (TAMP) efficiently. Our experimental results show that our framework significantly outperforms alternative optimization-based approaches for TAMP.
translated by 谷歌翻译
结肠镜检查是一种常规门诊手术,用于检查结肠和直肠的任何异常,包括息肉,憩室和结肠结构的狭窄。临床医生的大量时间用于在结肠镜检查过程中拍摄的快照,以维持医疗记录或进一步研究。自动化此步骤可以节省时间并提高流程的效率。在我们的工作中,我们收集了一个由专家注释的过程中的120个结肠镜检查视频和2416张快照的数据集。此外,我们开发了一种基于新颖的,视觉转化器的地标检测算法,该算法可以从结肠镜检查过程中鉴定出关键的解剖标志(阑尾孔,回肠瓣膜/盲肠地标和直肠翻新)。我们的算法在预处理过程中使用自适应伽马校正,以保持所有图像的一致亮度。然后,我们将视觉变压器用作特征提取主链和完全连接的基于网络的分类器头,将给定的框架分为四个类:三个地标或非地标框架。我们将视觉变压器(VIT-B/16)主链与RESNET-101和Convnext-B骨干进行了比较,这些骨干和Convnext-B骨干也接受了类似训练。我们报告了快照的测试数据集上的视觉变压器主链的精度为82%。
translated by 谷歌翻译
在视频分析中,背景模型具有许多应用,例如背景/前景分离,变更检测,异常检测,跟踪等。但是,尽管在静态相机捕获的视频中学习这种模型是一项公认的任务,但在移动相机背景模型(MCBM)的情况下,由于算法和可伸缩性挑战,成功率更加重要。由于相机运动而产生。因此,现有的MCBM在其范围和受支持的摄像头类型的限制中受到限制。这些障碍还阻碍了基于深度学习(DL)的端到端解决方案的这项无监督的任务。此外,现有的MCBM通常会在典型的大型全景图像或以在线方式的域名上建模背景。不幸的是,前者造成了几个问题,包括可扩展性差,而后者则阻止了对摄像机重新审视场景先前看到部分的案例的识别和利用。本文提出了一种称为DEEPMCBM的新方法,该方法消除了上述所有问题并实现最新结果。具体而言,首先,我们确定与一般和DL设置的视频帧联合对齐相关的困难。接下来,我们提出了一种新的联合一致性策略,使我们可以使用具有正则化的空间变压器网,也不是任何形式的专业化(且不差异)的初始化。再加上在不破坏的稳健中央矩(从关节对齐中获得)的自动编码器,这产生了一个无端到端的无端正规化MCBM,该MCBM支持广泛的摄像机运动并优雅地缩放。我们在各种视频上展示了DEEPMCBM的实用程序,包括超出其他方法范围的视频。我们的代码可在https://github.com/bgu-cs-vil/deepmcbm上找到。
translated by 谷歌翻译
评估患者结直肠癌的微卫星稳定性状态对于个性化治疗方案至关重要。最近,提出了卷积 - 神经网络(CNN)与转移学习方法结合使用,以规避传统的实验室测试,以确定苏木精和曙红染色的活检全幻灯片图像(WSI)的微卫星状态。但是,WSI的高分辨率实际上阻止了整个WSI的直接分类。当前方法通过先对WSI提取的小斑块进行分类,然后汇总补丁级分类徽标来推断患者级状态,从而绕过WSI高分辨率。这种方法限制了捕获位于高分辨率WSI数据的重要信息的能力。我们引入了一种有效的方法,通过对贴片嵌入的动量学习以及在这些嵌入组的组上培训患者级分类器,以利用WSI高分辨率信息。与直接的补丁级分类和患者水平聚合方法相比,我们的方法的准确性高达7.4 \%(AUC,$ 0.91 \ pm 0.01 $ vs. $ 0.85 \ $ 0.85 \ pm 0.04 $,p Value $ <0.01 $ )。我们的代码可以在https://github.com/technioncomputationalmrilab/coleroctal_cancer_ai上找到。
translated by 谷歌翻译
最近有很多不可能的结果表明,在与对抗对手的马尔可夫游戏中最小化的遗憾在统计学上和计算上是棘手的。然而,这些结果都没有排除在所有各方采用相同学习程序的假设下,遗憾最小化的可能性。在这项工作中,我们介绍了第一种(据我们所知)在通用马尔可夫游戏中学习的算法,该算法在所有代理商执行时提供了sublinear后悔保证。我们获得的边界是为了置换遗憾,因此,在此过程中,意味着融合了相关的平衡。我们的算法是分散的,计算上有效的,并且不需要代理之间的任何通信。我们的主要观察结果是,在马尔可夫游戏中通过策略优化的在线学习基本上减少了一种加权遗憾的最小化形式,而未知权重由代理商的策略顺序的路径长度确定。因此,控制路径长度会导致加权的遗憾目标,以提供足够的适应性算法提供统一的后悔保证。
translated by 谷歌翻译
虽然视觉和语言模型在视觉问题回答等任务上表现良好,但在基本的人类常识性推理技能方面,它们会挣扎。在这项工作中,我们介绍了Winogavil:在线游戏,以收集视觉和语言协会(例如,狼人到满月),用作评估最先进模型的动态基准。受欢迎的纸牌游戏代号的启发,Spymaster提供了与几个视觉候选者相关的文本提示,另一个玩家必须识别它们。人类玩家因创建对竞争对手AI模型而具有挑战性的联想而获得了回报,但仍然可以由其他人类玩家解决。我们使用游戏来收集3.5k实例,发现它们对人类的直观(> 90%的Jaccard索引),但对最先进的AI模型充满挑战,其中最佳模型(Vilt)的得分为52% ,成功的位置在视觉上是显着的。我们的分析以及我们从玩家那里收集的反馈表明,收集的关联需要多种推理技能,包括一般知识,常识,抽象等。我们发布数据集,代码和交互式游戏,旨在允许未来的数据收集,可用于开发具有更好关联能力的模型。
translated by 谷歌翻译