Deep Generative Vision as Approximate Bayesian Computation

Deep Generative Vision as Approximate Bayesian Computation
复制标题

深度生成视觉作为近似贝叶斯计算

DOI：
--
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
J. Tenenbaum
J. Tenenbaum
中科院分区：
--
文献类型：
--
作者：
Tejas D. Kulkarni;Ilker Yildirim;Pushmeet Kohli;W. Freiwald;J. Tenenbaum

文献摘要

被引文献

相似文献

逆图形的概率公式最近被提出用于各种2D和3D视觉问题[15，12，14，9]。这些方法以图形模拟器的形式表示视觉元素，该图形模拟器产生视觉场景的近似渲染。现有的方法要么是对像素数据建模，要么是手工制作的中间表示，如边缘图、超像素、轮廓等。然而，特征的选择会极大地影响推理质量和运行时间。最近，卷积神经网络(CNN)等深度学习技术在目标识别、场景像素标注等任务中表现出了令人印象深刻的性能，表明了基于CNN的特征的优越性。在这一发现的鼓舞下，我们测试了CNN与近似贝叶斯计算(ABC)相结合从单幅图像中反演出高维生成性逆图形模型的能力。我们成功地应用了概率近似MCMC算法的变体[21]，该算法使用CNN对两个现实世界问题的汇总统计进行量化：推断人类的3D姿势和从单幅图像生成人脸分析。计算机图形学似乎在为困难的图像合成问题设计解决方案方面取得了长足的进步。我们的实验表明，丰富的概率逆图形模型和深度学习方法的结合可以直接利用这种模拟器来解决难求逆问题。
Probabilistic formulations of inverse graphics have recently been proposed for a variety of 2D and 3D vision problems [15, 12, 14, 9]. These approaches represent visual elements in form of graphics simulators that produce approximate renderings of the visual scenes. Existing approaches either model pixel data or hand-crafted intermediate representations such as edge maps, super-pixels, silhouettes etc. However, the choice of features can drastically affect inference quality and run-time. Recently, deep learning techniques such as Convolutional Neural Networks (CNNs) have demonstrated impressive performance on various tasks such as object recognition and scene pixel labeling, suggesting the superiority of CNN-based features. Encouraged by this findings, we test the ability of CNNs in combination with Approximate Bayesian Computation (ABC) to invert high dimensional generative inverse graphics models from single images. We successfully applied a variant of the probabilistic approximate MCMC algorithm [21] which uses CNN to quantify summary statistics on two real world problems: inferring 3D pose of humans and generative face analysis from single images. Computer Graphics seems to be advancing at a great pace in terms of designing solutions for hard image synthesis problems. Our experiments indicate that the combination of rich probabilistic inverse graphics models and deep learning approaches could utilize such simulators directly to solve the hard inversion problem.