Media Analytics | Patents

MEDIA ANALYTICS

PROJECTS

PEOPLE

PUBLICATIONS

Patents

Face-Aware Person Re-identification System

Patent Number: 12,080,100
Xiang Yu; Masoud Faraki; Manmohan Chandraker; Yi-Hsuan Tsai; Yumin Suh
September 3, 2024

Multi-Domain Semantic Segmentation With Label Shifts

Patent Number:12,045,992
Masoud Faraki; Yumin Suh; Yi-Hsuan Tsai; Manmohan Chandraker; Dongwan Kim; Sparsh Garg
July 23, 2024

Simulating Diverse Long-Term Future Trajectories In Road Scenes

Patent Number: 12,005,892
Manmohan Chandraker; Sriram Nochur Narayanan
June 11, 2024

Monocular 3D Object Localization From Temporal Aggregation

Patent Number: 11,987,236
Pan Ji; Buyu Liu; Bingbing Zhuang; Xiangyu Chen; Manmohan Chandraker
May 21, 2024

Domain Generalized Margin Via Meta-Learning For Deep Face Recognition

Patent Number: 11,977,602
Ramin Moslemi; Masoud Faraki; Yi-Hsuan Tsai; Chang Liu; Xiang Yu; Manmohan Chandraker
May 7, 2024

Face Recognition from Unseen Domains via Learning of Semantic Features

Patent Number 11,947,626
Masoud Faraki; Yumin Suh; Yi-Hsuan Tsai; Manmohan Chandraker; Xiang Yu
April 2, 2024

Multi-Agent Trajectory Prediction

Patent Number: 11,816,901
Francesco Pittaluga; Manmohan Chandraker; Sriram Nochur Narayanan; Ramin Moslemi; Buyu Liu
November 14, 2023

Facial Recognition For Masked Individuals

Patent Number: 11,710,346
Xiang Yu; Ting Wang; Yi-Hsuan Tsai; Masoud Faraki; Ming-Fang Huang; Jian Fang; Yue Tian; Francesco Pittaluga; Gaurav Sharma; Yuheng Chen; Manmohan Chandraker
July 25, 2023

Joint Rolling Shutter Image Stitching and Rectification

Patent Number: 11,694,311
Bingbing Zhuang; Quoc-Huy Tran
July 04, 2023

Self-Supervised Cross-Video Temporal Difference Learning For Unsupervised Domain Adaptation

Patent Number: 11,676,370
Jinwoo Choi; Gaurav Sharma
June 13, 2023

Human Detection In Scenes

Patent Number: 11,610,420
Buyu Liu; Kihyuk Sohn; Jong-Chyi Su; Manmohan Chandraker; Yi-Hsuan Tsai
March 21, 2023

Domain Adaptation For Structured Output Via Disentangled Representations

Patent Number: 11,604,943
Kihyuk Sohn; Samuel Schulter; Manmohan Chandraker; Yi-Hsuan Tsai
March 14, 2023

Segmentation To Determine Lane Markings And Road Signs

Patent Number: 11,604,945
Yi-Hsuan Tsai; Manmohan Chandraker; Kihyuk Sohn; Buyu Liu; Jong-Chyi Su
March 14, 2023

Deep Face Recognition Based On Clustering Over Unlabeled Face Data

Patent Number: 11,600,113
Xiang Yu; Manmohan Chandraker; Aruni RoyChowdhury; Kihyuk Sohn
March 7, 2023

Joint Rolling Shutter Correction And Image Deblurring

Patent Number: 11,599,974
Quoc-Huy Tran; Manmohan Chandraker; Bingbing Zhuang; Pan Ji
March 7, 2023

Obstacle Detection In Road Scenes

Patent Number: 11,594,041
Kihyuk Sohn; Buyu Liu; Yi-Hsuan Tsai; Manmohan Chandraker; Jong-Chyi Su
February 28, 2023

Construction Zone Segmentation

Patent Number: 11,580,334
Manmohan Chandraker; Buyu Liu; Jong-Chyi Su; Kihyuk Sohn; Yi-Hsuan Tsai
February 14, 2023

Universal Feature Representation Learning For Face Recognition

Patent Number: 11,580,780
Yichun Shi; Xiang Yu; Manmohan Chandraker; Kihyuk Sohn
February 14, 2023

Learning To Simulate

Patent Number: 11,518,382
Nataniel Ruiz; Samuel Schulter; Manmohan Chandraker
December 6, 2022

Privacy-Preserving Visual Recognition Via Adversarial Learning

Patent Number: 11,520,923
Yi-Hsuan Tsai; Manmohan Chandraker; Kihyuk Sohn
December 6, 2022

Scene Attribute Annotation of Complex Road Typographies

Patent Number: 11,468,591
Samuel Schulter
October 11, 2022

Shuffle, Attend, and Adapt: Video Domain Adaptation by Clip Order Prediction and Clip Attention Alignment

Patent Number: 11,468,680
Gaurav Sharma; Samuel Schulter; Jinwoo Choi
October 11, 2022

Pseudo RGB-D for Self-Improving Monocular Slam and Depth Prediction

Patent Number: 11,468,585
Manmohan Chandraker; Quoc-Huy Tran; Pan Ji; Lokender Tiwari
October 11, 2022

Multi-Task Perception Network with Applications to Scene Understanding and Advanced Driver-Assistance System

Patent Number: 11,462,112
Buyu Liu; Paul Vernaza; Samuel Schulter; Manmohan Chandraker; Yi-Hsuan Tsai; Quoc-Huy Tran; Pan Ji
October 4, 2022

Parametric Top-View Representation of Complex Road Scenes

Patent Number: 11,455,813
Samuel Schulter; Buyu Liu; Manmohan Chandraker; Bingbing Zhuang
September 27, 2022

Parametric Top-View Representation Of Scenes

Patent Number: 11,373,067
Buyu Liu; Ziyan Wang; Samuel Schulter; Manmohan Chandraker
June 28, 2022

Self-Supervised Visual Odometry Framework Using Long-Term Modeling And Incremental Learning

Patent Number: 11,321,853
Pan Ji; Manmohan Chandraker; Yuliang Zou; Quoc-Huy Tran
May 3, 2022

Action Recognition System For Action Recognition In Unlabeled Videos With Domain Adversarial Learning And Knowledge Distillation

Patent Number: 11,314,993
Manmohan Chandraker; Kihyuk Sohn; Xiang Yu
April 26, 2022

Unsupervised Domain Adaptation for Video Classification

Patent Number: 11,301,716
Manmohan Chandraker; Jinwoo Choi; Gaurav SharApril 12, 2022

Face Spoofing Detection Using A Physical-Cue-Guided Multi-Source Multi-Channel Framework

Patent Number: 11,250,282
Junru Wu; Xiang Yu; Buyu Liu; Manmohan Chandraker
February 15, 2022

Human Action Recognition In Drone Videos

Patent Number: 11,250,573
Gaurav Sharma; Manmohan Chandraker; Jinwoo Choi
February 15, 2022

Image/Video Deblurring Using Convolutional Neural Networks With Applications To Sfm/Slam With Blurred Images/Videos

Patent Number: 11,222,409
Quoc-Huy Tran; Pan Ji; Manmohan Chandraker; Bingbing Zhuang
January 11, 2022

Object Detection With Training From Multiple Datasets

Patent Number: 11,222,238
Samuel Schulter; Yi-Hsuan Tsai; Xiangyun Zhao; Manmohan Chandraker; Gaurav SharJanuary 11, 2022

Attention And Warping Based Domain Adaptation For Videos

Patent Number: 11,222,210
Gaurav Sharma; Manmohan Chandraker
January 11, 2022

Traffic Prediction With Reparameterized Pushforward Policy For Autonomous Vehicles

Patent Number: 11,189,171
Nicholas Rhinehart; Paul Vernaza
November 30, 2021

Rolling Shutter Rectification In Images/Videos Using Convolutional Neural Networks With Applications To Sfm/Slam With Rolling Shutter Images/Videos

Patent Number: 11,132,586
Manmohan Chandraker; Quoc-Huy Tran; Pan Ji; Bingbing Zhuang
September 28, 2021

Recognizing Fine-Grained Objects In Surveillance Camera Images

Patent Number: 11,087,142
Shuyang Dai; Kihyuk Sohn; Yi-Hsuan Tsai; Manmohan Chandraker
August 10, 2021

Viewpoint Invariant Object Recognition By Synthesization And Domain Adaptation

Patent Number: 11,055,989
Luan Tran; Kihyuk Sohn; Manmohan Chandraker; Xiang Yu
July 6, 2021

Balancing Diversity And Precision Of Generative Models With Complementary Density Estimators

Patent Number: 11,049,265
Anqi Liu; Kihyuk Sohn; Nicholas Rhinehart; Paul Vernaza
June 29, 2021

Pose-Variant 3D Facial Attribute Generation

Patent Number: 10,991,145
Feng-Ju Chang; Xiang Yu; Manmohan Chandraker
April 27, 2021

Media Analytics News

A Parametric Top-View Representation of Complex Road Scenes

June 16, 2019

In this paper, we address the problem of inferring the layout of complex road scenes given a single camera as input. To achieve that, we first propose a novel parameterized model of road layouts in a top-view representation, which is not only intuitive for human visualization but also provides an interpretable interface for higher-level decision making. Moreover, the design of our top-view scene model allows for efficient sampling and thus generation of large-scale simulated data, which we leverage to train a deep neural network to infer our scene model's parameters. Specifically, our proposed training procedure uses supervised domain-adaptation techniques to incorporate both simulated as well as manually annotated data. Finally, we design a Conditional Random Field (CRF) that enforces coherent predictions for a single frame and encourages temporal smoothness among video frames. Experiments on two public data sets show that: (1) Our parametric top-view model is representative enough to describe complex road scenes; (2) The proposed method outperforms baselines trained on manually-annotated or simulated data only, thus getting the best of both; (3) Our CRF is able to generate temporally smoothed while semantically meaningful results.

Feature Transfer Learning for Face Recognition with Under-Represented Data

June 16, 2019

Despite the large volume of face recognition datasets, there is a significant portion of subjects, of which the samples are insufficient and thus under-represented. Ignoring such significant portion results in insufficient training data. Training with under-represented data leads to biased classifiers in conventionally-trained deep networks. In this paper, we propose a center-based feature transfer framework to augment the feature space of under-represented subjects from the regular subjects that have sufficiently diverse samples. A Gaussian prior of the variance is assumed across all subjects and the variance from regular ones are transferred to the under-represented ones. This encourages the under-represented distribution to be closer to the regular distribution. Further, an alternating training regimen is proposed to simultaneously achieve less biased classifiers and a more discriminative feature representation. We conduct ablative study to mimic the under-represented datasets by varying the portion of under-represented classes on the MS-Celeb-1M dataset. Advantageous results on LFW, IJB-A and MS-Celeb-1M demonstrate the effectiveness of our feature transfer and training strategy, compared to both general baselines and state-of-the-art methods. Moreover, our feature transfer successfully presents smooth visual interpolation, which conducts disentanglement to preserve identity of a class while augmenting its feature space with non-identity variations such as pose and lighting.

Gotta Adapt Em All: Joint Pixel and Feature-Level Domain Adaptation for Recognition in the Wild

June 16, 2019

Recent developments in deep domain adaptation have allowed knowledge transfer from a labeled source domain to an unlabeled target domain at the level of intermediate features or input pixels. We propose that advantages may be derived by combining them, in the form of different insights that lead to a novel design and complementary properties that result in better performance. At the feature level, inspired by insights from semi-supervised learning, we propose a classification-aware domain adversarial neural network that brings target examples into more classifiable regions of source domain. Next, we posit that computer vision insights are more amenable to injection at the pixel level. In particular, we use 3D geometry and image synthesis based on a generalized appearance flow to preserve identity across pose transformations, while using an attribute-conditioned CycleGAN to translate a single source into multiple target images that differ in lower-level properties such as lighting. Besides standard UDA benchmark, we validate on a novel and apt problem of car recognition in unlabeled surveillance images using labeled images from the web, handling explicitly specified, nameable factors of variation through pixel-level and implicit, unspecified factors through feature-level adaptation.

Media Analytics | Patents

Patents

Face-Aware Person Re-identification System

Multi-Domain Semantic Segmentation With Label Shifts

Simulating Diverse Long-Term Future Trajectories In Road Scenes

Monocular 3D Object Localization From Temporal Aggregation

Domain Generalized Margin Via Meta-Learning For Deep Face Recognition

Face Recognition from Unseen Domains via Learning of Semantic Features

Multi-Agent Trajectory Prediction

Facial Recognition For Masked Individuals

Joint Rolling Shutter Image Stitching and Rectification

Self-Supervised Cross-Video Temporal Difference Learning For Unsupervised Domain Adaptation

Human Detection In Scenes

Domain Adaptation For Structured Output Via Disentangled Representations

Segmentation To Determine Lane Markings And Road Signs

Deep Face Recognition Based On Clustering Over Unlabeled Face Data

Joint Rolling Shutter Correction And Image Deblurring

Obstacle Detection In Road Scenes

Construction Zone Segmentation

Universal Feature Representation Learning For Face Recognition

Learning To Simulate

Privacy-Preserving Visual Recognition Via Adversarial Learning

Scene Attribute Annotation of Complex Road Typographies

Shuffle, Attend, and Adapt: Video Domain Adaptation by Clip Order Prediction and Clip Attention Alignment

Pseudo RGB-D for Self-Improving Monocular Slam and Depth Prediction

Multi-Task Perception Network with Applications to Scene Understanding and Advanced Driver-Assistance System

Parametric Top-View Representation of Complex Road Scenes

Parametric Top-View Representation Of Scenes

Self-Supervised Visual Odometry Framework Using Long-Term Modeling And Incremental Learning

Action Recognition System For Action Recognition In Unlabeled Videos With Domain Adversarial Learning And Knowledge Distillation

Unsupervised Domain Adaptation for Video Classification

Face Spoofing Detection Using A Physical-Cue-Guided Multi-Source Multi-Channel Framework

Human Action Recognition In Drone Videos

Image/Video Deblurring Using Convolutional Neural Networks With Applications To Sfm/Slam With Blurred Images/Videos

Object Detection With Training From Multiple Datasets

Attention And Warping Based Domain Adaptation For Videos

Traffic Prediction With Reparameterized Pushforward Policy For Autonomous Vehicles

Rolling Shutter Rectification In Images/Videos Using Convolutional Neural Networks With Applications To Sfm/Slam With Rolling Shutter Images/Videos

Recognizing Fine-Grained Objects In Surveillance Camera Images

Viewpoint Invariant Object Recognition By Synthesization And Domain Adaptation

Balancing Diversity And Precision Of Generative Models With Complementary Density Estimators

Pose-Variant 3D Facial Attribute Generation

Media Analytics News

Contact Us

About Us

Our Pages

Read Our Blog Posts