To address the inference over the loopy human structure, our parser relies on a convolutional, message passing based approximation algorithm, which enjoys the advantages of iterative optimization and spatial information preservation
We showed that a proper graph-based spatio-temporal setup for pedestrian trajectory prediction improves over previous methods on several key aspects, including prediction error, computational time and number of parameters
This paper proposes steps towards this by inferring a rich representation of hands engaged in interaction method that includes: hand location, side, contact state, and a box around the object in contact
We present TailorNet, a neural model which predicts clothing deformation in 3D as a function of three factors: pose, shape and style, while retaining wrinkle detail
We have introduced the problem of human grasp prediction in RGB images and proposed GanHand, a generative model that 1) estimates the 3D pose of the objects in the scene; 2) predicts grasp types; and 3) refines a 3D hand mesh model
MANO Spectral Spiral GMM Spiral GMM, tune Spiral metrics, we report the F-score at a given threshold d which is the harmonic mean of precision and recall
Though more research is necessary to lower the financial cost and computational complexity of the NLOS imaging system described in this work, we believe that this preliminary work shows the remarkable potential for higher-level reasoning using NLOS imaging in the real-world
We have demonstrated that deep learning architectures that integrate combinatorial graph matching solvers perform well on deep graph matching benchmarks
Our experiments show that our non-linear optimization method is accurate enough to compute a training set of clothing images aligned with 3D mesh projections, from which we learn a direct mapping with a neural model
