no code implementations • 9 Sep 2023 • Sudeepta Mondal, Ganesh Sundaramoorthi
We propose a simple, efficient, and accurate method for detecting out-of-distribution (OOD) data for trained neural networks.
1 code implementation • 28 May 2023 • Huizong Yang, Yuxin Sun, Ganesh Sundaramoorthi, Anthony Yezzi
We show analytically that as the representation power of the network increases, the optimization approaches a partial differential equation (PDE) in the continuum limit that is unstable.
no code implementations • 4 Jun 2022 • Yuxin Sun, Dong Lao, Ganesh Sundaramoorthi, Anthony Yezzi
We discover restrained numerical instabilities in current training practices of deep networks with stochastic gradient descent (SGD).
no code implementations • NeurIPS Workshop DLDE 2021 • Naeemullah Khan, Angira Sharma, Philip Torr, Ganesh Sundaramoorthi
ST-DNN are deep networks formulated through the use of partial differential equations (PDE) to be defined on arbitrarily shaped regions.
no code implementations • NeurIPS Workshop DLDE 2021 • Yuxin Sun, Dong Lao, Ganesh Sundaramoorthi, Anthony Yezzi
We introduce a recently developed framework PDE Acceleration, which is a variational approach to accelerated optimization with partial differential equations (PDE), in the context of optimization of deep networks.
no code implementations • ICCV 2021 • Dong Lao, Peihao Zhu, Peter Wonka, Ganesh Sundaramoorthi
We consider the problem of filling in missing spatio-temporal regions of a video.
1 code implementation • 16 Jul 2021 • Angira Sharma, Naeemullah Khan, Muhammad Mubashar, Ganesh Sundaramoorthi, Philip Torr
For low-fidelity training data (incorrect class label) class-agnostic segmentation loss outperforms the state-of-the-art methods on salient object detection datasets by staggering margins of around 50%.
no code implementations • 16 Feb 2021 • Naeemullah Khan, Angira Sharma, Ganesh Sundaramoorthi, Philip H. S. Torr
We stack multiple PDE layers to generalize a deep CNN to arbitrary regions, and apply it to segmentation.
no code implementations • 1 Jan 2021 • Naeemullah Khan, Angira Sharma, Philip Torr, Ganesh Sundaramoorthi
We present Shape-Tailored Deep Neural Networks (ST-DNN).
1 code implementation • 28 Oct 2020 • Angira Sharma, Naeemullah Khan, Ganesh Sundaramoorthi, Philip Torr
For low-fidelity training data (incorrect class label) class-agnostic segmentation loss outperforms the state-of-the-art methods on salient object detection datasets by staggering margins of around 50%.
no code implementations • 25 Aug 2020 • Dong Lao, Peihao Zhu, Peter Wonka, Ganesh Sundaramoorthi
We introduce optimization methods for convolutional neural networks that can be used to improve existing gradient-based optimization in terms of generalization error.
1 code implementation • CVPR 2020 • Yanchao Yang, Dong Lao, Ganesh Sundaramoorthi, Stefano Soatto
We introduce two criteria to regularize the optimization involved in learning a classifier in a domain where no annotated data are available, leveraging annotated data in a different domain, a problem known as unsupervised domain adaptation.
no code implementations • 25 Nov 2019 • Ganesh Sundaramoorthi, Timothy E. Wang
This is shown to be invariant to small shifts, and preserves the efficiency of training.
1 code implementation • ICCV 2019 • Dong Lao, Ganesh Sundaramoorthi
We consider the problem of detecting objects, as they come into view, from videos in an online fashion.
no code implementations • NeurIPS 2018 • Ganesh Sundaramoorthi, Anthony Yezzi
Our approach evolves an infinite number of particles endowed with mass, represented as a mass density.
1 code implementation • ECCV 2018 • Dong Lao, Ganesh Sundaramoorthi
We consider the problem of inferring a layered representa-tion, its depth ordering and motion segmentation from a video in whichobjects may undergo 3D non-planar motion relative to the camera.
no code implementations • CVPR 2018 • Naeemullah Khan, Ganesh Sundaramoorthi
We formulate and optimize a joint optimization problem in the segmentation and descriptors to discriminate these base descriptors using the learned metric, equivalent to grouping learned descriptors.
no code implementations • 4 Apr 2018 • Ganesh Sundaramoorthi, Anthony Yezzi
We present a new class of optimization methods, valid for any optimization problem setup on the space of diffeomorphisms by generalizing Nesterov accelerated optimization to the manifold of diffeomorphisms.
no code implementations • 27 Nov 2017 • Anthony Yezzi, Ganesh Sundaramoorthi
Following the seminal work of Nesterov, accelerated optimization methods have been used to powerfully boost the performance of first-order, gradient-based parameter estimation in scenarios where second-order optimization strategies are either inapplicable or impractical.
no code implementations • CVPR 2017 • Dong Lao, Ganesh Sundaramoorthi
Our method is designed to detect the object(s) with minimum delay, i. e., frames after the object moves, constraining the false alarms.
no code implementations • CVPR 2017 • Naeemullah Khan, Byung-Woo Hong, Anthony Yezzi, Ganesh Sundaramoorthi
We formulate an energy for segmentation that is designed to have preference for segmenting the coarse over fine structure of the image, without smoothing across boundaries of regions.
no code implementations • 30 Apr 2017 • Marei Algarni, Ganesh Sundaramoorthi
We present SurfCut, an algorithm for extracting a smooth, simple surface with an unknown 3D curve boundary from a noisy 3D image and a seed point.
no code implementations • 24 May 2016 • Dong Lao, Ganesh Sundaramoorthi
We present a general framework and method for simultaneous detection and segmentation of an object in a video that moves (or comes into view of the camera) at some unknown time in the video.
no code implementations • 24 Mar 2016 • Ganesh Sundaramoorthi, Naeemullah Khan, Byung-Woo Hong
We formulate a general energy and method for segmentation that is designed to have preference for segmenting the coarse structure over the fine structure of the data, without smoothing across boundaries of regions.
no code implementations • ICCV 2015 • Yanchao Yang, Ganesh Sundaramoorthi, Stefano Soatto
We propose a method to detect disocclusion in video sequences of three-dimensional scenes and to partition the disoccluded regions into objects, defined by coherent deformation corresponding to surfaces in the scene.
no code implementations • CVPR 2015 • Yanchao Yang, Zhaojin Lu, Ganesh Sundaramoorthi
We present a new approach to wide baseline matching.
no code implementations • CVPR 2015 • Naeemullah Khan, Marei Algarni, Anthony Yezzi, Ganesh Sundaramoorthi
Given a region of arbitrary shape in an image, these descriptors are formed from shape-dependent scale spaces of oriented gradients.
no code implementations • CVPR 2014 • Ganesh Sundaramoorthi, Byung-Woo Hong
We derive an easy-to-implement and efficient algorithm for solving multi-label image partitioning problems in the form of the problem addressed by Region Competition.
no code implementations • 6 Feb 2014 • Omar Arif, Ganesh Sundaramoorthi, Byung-Woo Hong, Anthony Yezzi
We illustrate the use of this motion estimation scheme in propagating a segmentation across frames and show that it leads to more accurate segmentation than traditional motion estimation that does not make use of physical constraints.
no code implementations • CVPR 2013 • Byung-Woo Hong, Zhaojin Lu, Ganesh Sundaramoorthi
The advantage of this statistical model is that the underlying variables: the labels and the functions are less coupled than in the original formulation, and the labels can be computed from the functions with more global updates.
no code implementations • 21 Aug 2012 • Yanchao Yang, Ganesh Sundaramoorthi
In cases of 3D object motion and viewpoint change, self-occlusions and dis-occlusions of the object are prominent, and current methods employing joint shape and appearance models are unable to adapt to new shape and appearance information, leading to inaccurate shape detection.