Search Results for author: Jia Deng

Found 74 papers, 39 papers with code

Siamese Masked Autoencoders

no code implementations23 May 2023 Agrim Gupta, Jiajun Wu, Jia Deng, Li Fei-Fei

Establishing correspondence between images or scenes is a significant challenge in computer vision, especially given occlusions, viewpoint changes, and varying object appearances.

Data Augmentation Semantic Segmentation +2

Label-Free Synthetic Pretraining of Object Detectors

1 code implementation8 Aug 2022 Hei Law, Jia Deng

Our "SOLID" approach consists of two main components: (1) generating synthetic images using a collection of unlabelled 3D models with optimized scene arrangement; (2) pretraining an object detector on "instance detection" task - given a query image depicting an object, detecting all instances of the exact same object in a target image.

Deep Patch Visual Odometry

1 code implementation8 Aug 2022 Zachary Teed, Lahav Lipson, Jia Deng

DPVO disproves this assumption, showing that it is possible to get the best accuracy and efficiency by exploiting the advantages of sparse patch-based matching over dense flow.

Monocular Visual Odometry

Generating Natural Language Proofs with Verifier-Guided Search

1 code implementation25 May 2022 Kaiyu Yang, Jia Deng, Danqi Chen

In this paper, we present a novel stepwise method, NLProofS (Natural Language Proof Search), which learns to generate relevant steps conditioning on the hypothesis.

View Synthesis with Sculpted Neural Points

1 code implementation12 May 2022 Yiming Zuo, Jia Deng

In this work, we propose a new approach that performs view synthesis using point clouds.

Multiview Stereo with Cascaded Epipolar RAFT

1 code implementation9 May 2022 Zeyu Ma, Zachary Teed, Jia Deng

CER-MVS is significantly different from prior work in multiview stereo.

Optical Flow Estimation

Coupled Iterative Refinement for 6D Multi-Object Pose Estimation

1 code implementation CVPR 2022 Lahav Lipson, Zachary Teed, Ankit Goyal, Jia Deng

We propose a new approach to 6D object pose estimation which consists of an end-to-end differentiable architecture that makes use of geometric knowledge.

6D Pose Estimation using RGB

IFOR: Iterative Flow Minimization for Robotic Object Rearrangement

no code implementations CVPR 2022 Ankit Goyal, Arsalan Mousavian, Chris Paxton, Yu-Wei Chao, Brian Okorn, Jia Deng, Dieter Fox

Accurate object rearrangement from vision is a crucial problem for a wide variety of real-world robotics applications in unstructured environments.

Optical Flow Estimation

Learning Symbolic Rules for Reasoning in Quasi-Natural Language

2 code implementations23 Nov 2021 Kaiyu Yang, Jia Deng

In this work, we ask how we can build a rule-based system that can reason with natural language input but without the manual construction of rules.

Automated Theorem Proving Formal Logic +1

Non-deep Networks

4 code implementations14 Oct 2021 Ankit Goyal, Alexey Bochkovskiy, Jia Deng, Vladlen Koltun

This begs the question -- is it possible to build high-performing "non-deep" neural networks?

Image Classification Real-Time Object Detection

Dynamically Grown Generative Adversarial Networks

no code implementations16 Jun 2021 Lanlan Liu, Yuting Zhang, Jia Deng, Stefano Soatto

Recent work introduced progressive network growing as a promising way to ease the training for large GANs, but the model design and architecture-growing strategy still remain under-explored and needs manual design for different image data.

Image Generation

Revisiting Point Cloud Shape Classification with a Simple and Effective Baseline

3 code implementations9 Jun 2021 Ankit Goyal, Hei Law, Bowei Liu, Alejandro Newell, Jia Deng

It also outperforms state-of-the-art methods on ScanObjectNN, a real-world point cloud benchmark, and demonstrates better cross-dataset generalization.

Point Cloud Classification

Tangent Space Backpropagation for 3D Transformation Groups

1 code implementation CVPR 2021 Zachary Teed, Jia Deng

We address the problem of performing backpropagation for computation graphs involving 3D transformation groups SO(3), SE(3), and Sim(3).

A Study of Face Obfuscation in ImageNet

1 code implementation10 Mar 2021 Kaiyu Yang, Jacqueline Yau, Li Fei-Fei, Jia Deng, Olga Russakovsky

In this paper, we explore the effects of face obfuscation on the popular ImageNet challenge visual recognition benchmark.

object-detection Object Detection +3

HYPE-C: Evaluating Image Completion Models Through Standardized Crowdsourcing

no code implementations1 Jan 2021 Emily Walters, Weifeng Chen, Jia Deng

Recent work has proposed the use of human evaluation for image synthesis models, allowing for a reliable method to evaluate the visual quality of generated images.

Image Generation

Revisiting Point Cloud Classification with a Simple and Effective Baseline

2 code implementations1 Jan 2021 Ankit Goyal, Hei Law, Bowei Liu, Alejandro Newell, Jia Deng

It also outperforms state-of-the-art methods on ScanObjectNN, a real-world point cloud benchmark, and demonstrates better cross-dataset generalization.

3D Point Cloud Classification Classification +2

Rel3D: A Minimally Contrastive Benchmark for Grounding Spatial Relations in 3D

2 code implementations NeurIPS 2020 Ankit Goyal, Kaiyu Yang, Dawei Yang, Jia Deng

The 3D scenes in our dataset come in minimally contrastive pairs: two scenes in a pair are almost identical, but a spatial relation holds in one and fails in the other.

Spatial Relation Recognition

RAFT-3D: Scene Flow using Rigid-Motion Embeddings

1 code implementation CVPR 2021 Zachary Teed, Jia Deng

We address the problem of scene flow: given a pair of stereo or RGB-D video frames, estimate pixelwise 3D motion.

Optical Flow Estimation Scene Flow Estimation

A Unified Framework of Surrogate Loss by Refactoring and Interpolation

1 code implementation ECCV 2020 Lanlan Liu, Mingzhe Wang, Jia Deng

We introduce UniLoss, a unified framework to generate surrogate losses for training deep networks with gradient descent, reducing the amount of manual design of task-specific surrogate losses.

OASIS: A Large-Scale Dataset for Single Image 3D in the Wild

no code implementations CVPR 2020 Weifeng Chen, Shengyi Qian, David Fan, Noriyuki Kojima, Max Hamilton, Jia Deng

Single-view 3D is the task of recovering 3D properties such as depth and surface normals from a single image.

PackIt: A Virtual Environment for Geometric Planning

1 code implementation ICML 2020 Ankit Goyal, Jia Deng

The ability to jointly understand the geometry of objects and plan actions for manipulating them is crucial for intelligent agents.

Robot Task Planning

How Useful is Self-Supervised Pretraining for Visual Tasks?

2 code implementations CVPR 2020 Alejandro Newell, Jia Deng

We investigate what factors may play a role in the utility of these pretraining methods for practitioners.

RAFT: Recurrent All-Pairs Field Transforms for Optical Flow

9 code implementations ECCV 2020 Zachary Teed, Jia Deng

RAFT extracts per-pixel features, builds multi-scale 4D correlation volumes for all pairs of pixels, and iteratively updates a flow field through a recurrent unit that performs lookups on the correlation volumes.

Optical Flow Estimation

Compositional Temporal Visual Grounding of Natural Language Event Descriptions

no code implementations4 Dec 2019 Jonathan C. Stroud, Ryan McCaffrey, Rada Mihalcea, Jia Deng, Olga Russakovsky

Temporal grounding entails establishing a correspondence between natural language event descriptions and their visual depictions.

Visual Grounding

Representing Movie Characters in Dialogues

no code implementations CONLL 2019 Mahmoud Azab, Noriyuki Kojima, Jia Deng, Rada Mihalcea

We introduce a new embedding model to represent movie characters and their interactions in a dialogue by encoding in the same representation the language used by these characters as well as information about the other participants in the dialogue.

Question Answering Relation Classification +1

Generative Modeling for Small-Data Object Detection

1 code implementation ICCV 2019 Lanlan Liu, Michael Muelly, Jia Deng, Tomas Pfister, Li-Jia Li

This paper explores object detection in the small data regime, where only a limited number of annotated bounding boxes are available due to data rarity and annotation expense.

object-detection Object Detection +3

Feature Partitioning for Efficient Multi-Task Architectures

no code implementations ICLR 2020 Alejandro Newell, Lu Jiang, Chong Wang, Li-Jia Li, Jia Deng

Multi-task learning holds the promise of less data, parameters, and time than training of separate models.

Multi-Task Learning

To Learn or Not to Learn: Analyzing the Role of Learning for Navigation in Virtual Environments

no code implementations26 Jul 2019 Noriyuki Kojima, Jia Deng

In this paper we compare learning-based methods and classical methods for navigation in virtual environments.


Learning to Generate Synthetic 3D Training Data through Hybrid Gradient

no code implementations29 Jun 2019 Dawei Yang, Jia Deng

We parametrize the design decisions as a real vector, and combine the approximate gradient and the analytical gradient to obtain the hybrid gradient of the network performance with respect to this vector.

Learning to Prove Theorems via Interacting with Proof Assistants

1 code implementation21 May 2019 Kaiyu Yang, Jia Deng

Proof assistants offer a formalism that resembles human mathematical reasoning, representing theorems in higher-order logic and proofs as high-level tactics.

Automated Theorem Proving Mathematical Reasoning

CornerNet-Lite: Efficient Keypoint Based Object Detection

6 code implementations18 Apr 2019 Hei Law, Yun Teng, Olga Russakovsky, Jia Deng

Together these two variants address the two critical use cases in efficient object detection: improving efficiency without sacrificing accuracy, and improving accuracy at real-time efficiency.

object-detection Real-Time Object Detection

D3D: Distilled 3D Networks for Video Action Recognition

1 code implementation19 Dec 2018 Jonathan C. Stroud, David A. Ross, Chen Sun, Jia Deng, Rahul Sukthankar

State-of-the-art methods for video action recognition commonly use an ensemble of two networks: the spatial stream, which takes RGB frames as input, and the temporal stream, which takes optical flow as input.

Action Classification Action Recognition +2

MeshAdv: Adversarial Meshes for Visual Recognition

no code implementations CVPR 2019 Chaowei Xiao, Dawei Yang, Bo Li, Jia Deng, Mingyan Liu

Highly expressive models such as deep neural networks (DNNs) have been widely applied to various applications.

Speaker Naming in Movies

no code implementations NAACL 2018 Mahmoud Azab, Mingzhe Wang, Max Smith, Noriyuki Kojima, Jia Deng, Rada Mihalcea

We propose a new model for speaker naming in movies that leverages visual, textual, and acoustic modalities in an unified optimization framework.

Rethinking Numerical Representations for Deep Neural Networks

no code implementations7 Aug 2018 Parker Hill, Babak Zamirai, Shengshuo Lu, Yu-Wei Chao, Michael Laurenzano, Mehrzad Samadi, Marios Papaefthymiou, Scott Mahlke, Thomas Wenisch, Jia Deng, Lingjia Tang, Jason Mars

With ever-increasing computational demand for deep learning, it is critical to investigate the implications of the numeric representation and precision of DNN model weights and activations on computational efficiency.

CornerNet: Detecting Objects as Paired Keypoints

5 code implementations ECCV 2018 Hei Law, Jia Deng

We propose CornerNet, a new approach to object detection where we detect an object bounding box as a pair of keypoints, the top-left corner and the bottom-right corner, using a single convolution neural network.

object-detection Object Detection

Decorrelated Batch Normalization

6 code implementations CVPR 2018 Lei Huang, Dawei Yang, Bo Lang, Jia Deng

Batch Normalization (BN) is capable of accelerating the training of deep models by centering and scaling activations within mini-batches.

Shape from Shading through Shape Evolution

no code implementations CVPR 2018 Dawei Yang, Jia Deng

The evolution generates better shapes guided by the network training, while the training improves by using the evolved shapes.

Premise Selection for Theorem Proving by Deep Graph Embedding

1 code implementation NeurIPS 2017 Mingzhe Wang, Yihe Tang, Jian Wang, Jia Deng

We propose a deep learning-based approach to the problem of premise selection: selecting mathematical statements relevant for proving a given conjecture.

Automated Theorem Proving General Classification +1

Fine-Grained Car Detection for Visual Census Estimation

no code implementations7 Sep 2017 Timnit Gebru, Jonathan Krause, Yi-Lun Wang, Duyun Chen, Jia Deng, Li Fei-Fei

In this work, we leverage the ubiquity of Google Street View images and develop a computer vision pipeline to predict income, per capita carbon emission, crime rates and other city attributes from a single source of publicly available visual data.

Scalable Annotation of Fine-Grained Categories Without Experts

no code implementations7 Sep 2017 Timnit Gebru, Jonathan Krause, Jia Deng, Li Fei-Fei

We present a crowdsourcing workflow to collect image annotations for visually similar synthetic categories without requiring experts.

Temporal Action Localization by Structured Maximal Sums

no code implementations CVPR 2017 Zehuan Yuan, Jonathan C. Stroud, Tong Lu, Jia Deng

We pose action localization as a structured prediction over arbitrary-length temporal windows, where each window is scored as the sum of frame-wise classification scores.

Action Detection General Classification +2

Surface Normals in the Wild

no code implementations ICCV 2017 Weifeng Chen, Donglai Xiang, Jia Deng

We study the problem of single-image depth estimation for images in the wild.

Depth Estimation

Using Deep Learning and Google Street View to Estimate the Demographic Makeup of the US

no code implementations22 Feb 2017 Timnit Gebru, Jonathan Krause, Yi-Lun Wang, Duyun Chen, Jia Deng, Erez Lieberman Aiden, Li Fei-Fei

The United States spends more than $1B each year on initiatives such as the American Community Survey (ACS), a labor-intensive door-to-door study that measures statistics relating to race, gender, education, occupation, unemployment, and other demographic factors.

Learning to Detect Human-Object Interactions

no code implementations17 Feb 2017 Yu-Wei Chao, Yunfan Liu, Xieyang Liu, Huayi Zeng, Jia Deng

We study the problem of detecting human-object interactions (HOI) in static images, defined as predicting a human and an object bounding box with an interaction class label that connects them.

General Classification Human-Object Interaction Detection

Dynamic Deep Neural Networks: Optimizing Accuracy-Efficiency Trade-offs by Selective Execution

no code implementations2 Jan 2017 Lanlan Liu, Jia Deng

We introduce Dynamic Deep Neural Networks (D2NN), a new type of feed-forward deep neural network that allows selective execution.

Image Classification

Single-Image Depth Perception in the Wild

4 code implementations NeurIPS 2016 Weifeng Chen, Zhao Fu, Dawei Yang, Jia Deng

This paper studies single-image depth perception in the wild, i. e., recovering depth from a single image taken in unconstrained settings.

Depth Estimation

Stacked Hourglass Networks for Human Pose Estimation

41 code implementations22 Mar 2016 Alejandro Newell, Kaiyu Yang, Jia Deng

This work introduces a novel convolutional network architecture for the task of human pose estimation.

Pose Estimation

HICO: A Benchmark for Recognizing Human-Object Interactions in Images

no code implementations ICCV 2015 Yu-Wei Chao, Zhan Wang, Yugeng He, Jiaxuan Wang, Jia Deng

We introduce a new benchmark "Humans Interacting with Common Objects" (HICO) for recognizing human-object interactions (HOI).

Human-Object Interaction Detection

Mining Semantic Affordances of Visual Object Categories

no code implementations CVPR 2015 Yu-Wei Chao, Zhan Wang, Rada Mihalcea, Jia Deng

In this paper we introduce the new problem of mining the knowledge of semantic affordance: given an object, determining whether an action can be performed on it.

Collaborative Filtering

Probabilistic Label Relation Graphs with Ising Models

no code implementations ICCV 2015 Nan Ding, Jia Deng, Kevin Murphy, Hartmut Neven

In this paper, we extend the HEX model to allow for soft or probabilistic relations between labels, which is useful when there is uncertainty about the relationship between two labels (e. g., an antelope is "sort of" furry, but not to the same degree as a grizzly bear).

General Classification

ImageNet Large Scale Visual Recognition Challenge

12 code implementations1 Sep 2014 Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, Sanjeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, Alexander C. Berg, Li Fei-Fei

The ImageNet Large Scale Visual Recognition Challenge is a benchmark in object category classification and detection on hundreds of object categories and millions of images.

General Classification Image Classification +3

Fine-Grained Crowdsourcing for Fine-Grained Recognition

no code implementations CVPR 2013 Jia Deng, Jonathan Krause, Li Fei-Fei

In this work, we include humans in the loop to help computers select discriminative features.

feature selection

Cannot find the paper you are looking for? You can Submit a new open access paper.