Search Results for author: Hengduo Li

Found 12 papers, 3 papers with code

AdaViT: Adaptive Vision Transformers for Efficient Image Recognition

no code implementations30 Nov 2021 Lingchen Meng, Hengduo Li, Bor-Chun Chen, Shiyi Lan, Zuxuan Wu, Yu-Gang Jiang, Ser-Nam Lim

To this end, we introduce AdaViT, an adaptive computation framework that learns to derive usage policies on which patches, self-attention heads and transformer blocks to use throughout the backbone on a per-input basis, aiming to improve inference efficiency of vision transformers with a minimal drop of accuracy for image recognition.

Efficient Video Transformers with Spatial-Temporal Token Selection

no code implementations23 Nov 2021 Junke Wang, Xitong Yang, Hengduo Li, Zuxuan Wu, Yu-Gang Jiang

Video transformers have achieved impressive results on major video recognition benchmarks, however they suffer from high computational cost.

Video Recognition

Rethinking Pseudo Labels for Semi-Supervised Object Detection

no code implementations1 Jun 2021 Hengduo Li, Zuxuan Wu, Abhinav Shrivastava, Larry S. Davis

In this paper, we introduce certainty-aware pseudo labels tailored for object detection, which can effectively estimate the classification and localization quality of derived pseudo labels.

Classification Image Classification +2

HMS: Hierarchical Modality Selection for Efficient Video Recognition

no code implementations20 Apr 2021 Zejia Weng, Zuxuan Wu, Hengduo Li, Yu-Gang Jiang

Conventional video recognition pipelines typically fuse multimodal features for improved performance.

Video Recognition

Improving the Tightness of Convex Relaxation Bounds for Training Certifiably Robust Classifiers

no code implementations22 Feb 2020 Chen Zhu, Renkun Ni, Ping-Yeh Chiang, Hengduo Li, Furong Huang, Tom Goldstein

Convex relaxations are effective for training and certifying neural networks against norm-bounded adversarial attacks, but they leave a large gap between certifiable and empirical robustness.

Learning from Noisy Anchors for One-stage Object Detection

1 code implementation CVPR 2020 Hengduo Li, Zuxuan Wu, Chen Zhu, Caiming Xiong, Richard Socher, Larry S. Davis

State-of-the-art object detectors rely on regressing and classifying an extensive list of possible anchors, which are divided into positive and negative samples based on their intersection-over-union (IoU) with corresponding groundtruth objects.

Classification General Classification +1

Improved Training of Certifiably Robust Models

no code implementations25 Sep 2019 Chen Zhu, Renkun Ni, Ping-Yeh Chiang, Hengduo Li, Furong Huang, Tom Goldstein

Convex relaxations are effective for training and certifying neural networks against norm-bounded adversarial attacks, but they leave a large gap between certifiable and empirical (PGD) robustness.

Transferable Clean-Label Poisoning Attacks on Deep Neural Nets

1 code implementation15 May 2019 Chen Zhu, W. Ronny Huang, Ali Shafahi, Hengduo Li, Gavin Taylor, Christoph Studer, Tom Goldstein

Clean-label poisoning attacks inject innocuous looking (and "correctly" labeled) poison images into training data, causing a model to misclassify a targeted image after being trained on this data.

Transfer Learning

An Analysis of Pre-Training on Object Detection

no code implementations11 Apr 2019 Hengduo Li, Bharat Singh, Mahyar Najibi, Zuxuan Wu, Larry S. Davis

We analyze how well their features generalize to tasks like image classification, semantic segmentation and object detection on small datasets like PASCAL-VOC, Caltech-256, SUN-397, Flowers-102 etc.

Classification General Classification +3

R-FCN-3000 at 30fps: Decoupling Detection and Classification

2 code implementations CVPR 2018 Bharat Singh, Hengduo Li, Abhishek Sharma, Larry S. Davis

Our approach is a modification of the R-FCN architecture in which position-sensitive filters are shared across different object classes for performing localization.

Classification General Classification

Multi-Glimpse LSTM with Color-Depth Feature Fusion for Human Detection

no code implementations3 Nov 2017 Hengduo Li, Jun Liu, Guyue Zhang, Yuan Gao, Yirui Wu

In this paper, we propose a new Multi-Glimpse LSTM (MG-LSTM) network, in which multi-scale contextual information is sequentially integrated to promote the human detection performance.

Human Detection

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