Y
Yuwen Xiong
Researcher at University of Toronto
Publications - 38
Citations - 7038
Yuwen Xiong is an academic researcher from University of Toronto. The author has contributed to research in topics: Segmentation & Object detection. The author has an hindex of 18, co-authored 36 publications receiving 4081 citations. Previous affiliations of Yuwen Xiong include Uber & Microsoft.
Papers
More filters
Posted Content
Discrete Residual Flow for Probabilistic Pedestrian Behavior Prediction
TL;DR: This work proposes the discrete residual flow network (DRF-Net), a convolutional neural network for human motion prediction that captures the uncertainty inherent in long-range motion forecasting and effectively captures multimodal posteriors over future human motion by predicting and updating a discretized distribution over spatial locations.
Posted Content
LoCo: Local Contrastive Representation Learning.
TL;DR: By overlapping local blocks stacking on top of each other, this work effectively increases the decoder depth and allow upper blocks to implicitly send feedbacks to lower blocks, which closes the performance gap between local learning and end-to-end contrastive learning algorithms for the first time.
Posted Content
DMM-Net: Differentiable Mask-Matching Network for Video Object Segmentation.
TL;DR: Differentiable Mask Matching Network (DMM-Net) as discussed by the authors proposes a differentiable mask-matching layer by unrolling a projected gradient descent algorithm in which the projection exploits the Dykstra's algorithm.
Posted Content
PolyTransform: Deep Polygon Transformer for Instance Segmentation
TL;DR: PolyTransform as discussed by the authors uses a segmentation network to generate instance masks and then converts the masks into a set of polygons that are then fed to a deforming network that transforms the polygons such that they better fit the object boundaries.
Posted Content
UPSNet: A Unified Panoptic Segmentation Network
TL;DR: A parameter-free panoptic head is introduced which solves thepanoptic segmentation via pixel-wise classification and first leverages the logits from the previous two heads and then innovatively expands the representation for enabling prediction of an extra unknown class which helps better resolving the conflicts between semantic and instance segmentation.