no code implementations • 5 Dec 2023 • Shao-Yu Chang, Hwann-Tzong Chen, Tyng-Luh Liu
Despite the success in image editing, diffusion models still encounter significant hindrances when it comes to video editing due to the challenge of maintaining spatiotemporal consistency in the object's appearance across frames.
1 code implementation • ICML 2023 • Yu-Min Chu, Chieh Liu, Ting-I Hsieh, Hwann-Tzong Chen, Tyng-Luh Liu
We present a shape-guided expert-learning framework to tackle the problem of unsupervised 3D anomaly detection.
Ranked #1 on 3D Anomaly Detection and Segmentation on MVTEC 3D-AD
3D Anomaly Detection 3D Anomaly Detection and Segmentation +2
no code implementations • ICCV 2023 • Cheng-Yao Hong, Yu-Ying Chou, Tyng-Luh Liu
The proposed attention discriminant sampling (ADS) starts by efficiently decomposing a given point cloud into clusters to implicitly encode its structural and geometric relatedness among points.
1 code implementation • ECCV 2022 2022 • Jhih-Ciang Wu*, He-Yen Hsieh*, Ding-Jie Chen, Chiou-Shann Fuh, Tyng-Luh Liu
Video anomaly detection (VAD) aims at localizing unexpected actions or activities in a video sequence.
Anomaly Detection In Surveillance Videos Self-Supervised Learning
no code implementations • CVPR 2022 • Wen-Li Wei, Jen-Chun Lin, Tyng-Luh Liu, Hong-Yuan Mark Liao
To address this problem, we propose a motion pose and shape network (MPS-Net) to effectively capture humans in motion to estimate accurate and temporally coherent 3D human pose and shape from a video.
Ranked #52 on 3D Human Pose Estimation on 3DPW
no code implementations • 23 Dec 2021 • Ta-Ying Cheng, Hsuan-ru Yang, Niki Trigoni, Hwann-Tzong Chen, Tyng-Luh Liu
We present a pose adaptive few-shot learning procedure and a two-stage data interpolation regularization, termed Pose Adaptive Dual Mixup (PADMix), for single-image 3D reconstruction.
4 code implementations • 13 Oct 2021 • Chun-Hsiao Yeh, Cheng-Yao Hong, Yen-Chi Hsu, Tyng-Luh Liu, Yubei Chen, Yann Lecun
Further, DCL can be combined with the SOTA contrastive learning method, NNCLR, to achieve 72. 3% ImageNet-1K top-1 accuracy with 512 batch size in 400 epochs, which represents a new SOTA in contrastive learning.
no code implementations • 29 Sep 2021 • Bing-Jhang Lin, Ding-Jie Chen, He-Yen Hsieh, Tyng-Luh Liu
We comprehensively identify the missing neighborhood relationships issue of conventional embedding and propose a novel approach, termed as Graph Local Embedding (GLE), to deep metric learning.
no code implementations • CVPR 2021 • Ding-Jie Chen, He-Yen Hsieh, Tyng-Luh Liu
One-shot object detection tackles a challenging task that aims at identifying within a target image all object instances of the same class, implied by a query image patch.
1 code implementation • ICLR 2021 • Yu-Ying Chou, Hsuan-Tien Lin, Tyng-Luh Liu
In addition, to break the limit of training with images only from seen classes, we design a generative scheme to simultaneously generate virtual class labels and their visual features by sampling and interpolating over seen counterparts.
no code implementations • 1 Jan 2021 • He-Yen Hsieh, Ding-Jie Chen, Tung-Ying Lee, Tyng-Luh Liu
The task of temporal action proposal generation (TAPG) aims to provide high-quality video segments, i. e., proposals that potentially contain action events.
no code implementations • 1 Jan 2021 • Yen-Chi Hsu, Cheng-Yao Hong, Wan-Cyuan Fan, Ding-Jie Chen, Ming-Sui Lee, Davi Geiger, Tyng-Luh Liu
The Fine-Grained Visual Classification (FGVC) problem is notably characterized by two intriguing properties, significant inter-class similarity and intra-class variations, which cause learning an effective FGVC classifier a challenging task.
no code implementations • ICCV 2021 • Jhih-Ciang Wu, Ding-Jie Chen, Chiou-Shann Fuh, Tyng-Luh Liu
Anomaly detection (AD) aims to address the task of classification or localization of image anomalies.
1 code implementation • 20 Jul 2020 • Shih-Hung Liu, Shang-Yi Yu, Shao-Chi Wu, Hwann-Tzong Chen, Tyng-Luh Liu
This paper presents a novel method for instance segmentation of 3D point clouds.
Ranked #9 on 3D Instance Segmentation on S3DIS (mPrec metric)
no code implementations • ECCV 2020 • Hsien-Tzu Cheng, Chun-Fu Yeh, Po-Chen Kuo, Andy Wei, Keng-Chi Liu, Mong-Chi Ko, Kuan-Hua Chao, Yu-Ching Peng, Tyng-Luh Liu
Following this similarity learning, our similarity ensemble merges similar patches' ensembled predictions as the pseudo-label of a given patch to counteract its noisy label.
no code implementations • 24 Apr 2020 • Chun-Fu Yeh, Hsien-Tzu Cheng, Andy Wei, Hsin-Ming Chen, Po-Chen Kuo, Keng-Chi Liu, Mong-Chi Ko, Ray-Jade Chen, Po-Chang Lee, Jen-Hsiang Chuang, Chi-Mai Chen, Yi-Chang Chen, Wen-Jeng Lee, Ning Chien, Jo-Yu Chen, Yu-Sen Huang, Yu-Chien Chang, Yu-Cheng Huang, Nai-Kuan Chou, Kuan-Hua Chao, Yi-Chin Tu, Yeun-Chung Chang, Tyng-Luh Liu
We introduce a comprehensive screening platform for the COVID-19 (a. k. a., SARS-CoV-2) pneumonia.
2 code implementations • NeurIPS 2019 • Ting-I Hsieh, Yi-Chen Lo, Hwann-Tzong Chen, Tyng-Luh Liu
This paper aims to tackle the challenging problem of one-shot object detection.
Ranked #3 on One-Shot Object Detection on MS COCO
no code implementations • None 2019 • Yen-Chi Hsu, Cheng-Yao Hong, Ding-Jie Chen, Ming-Sui Lee, Davi Geiger, Tyng-Luh Liu
We introduce a regularization concept based on the proposed Batch Confusion Norm (BCN) to address Fine-Grained Visual Classification (FGVC).
Ranked #17 on Fine-Grained Image Classification on FGVC Aircraft
no code implementations • 28 Oct 2019 • Yen-Chi Hsu, Cheng-Yao Hong, Wan-Cyuan Fan, Ming-Sui Lee, Davi Geiger, Tyng-Luh Liu
With the development of deep learning, standard classification problems have achieved good results.
Fine-Grained Image Classification Fine-Grained Visual Recognition
no code implementations • 6 Apr 2019 • Chih-Yao Chiu, Hwann-Tzong Chen, Tyng-Luh Liu
This paper describes a channel-selection approach for simplifying deep neural networks.
3 code implementations • 20 Dec 2018 • Ding-Jie Chen, Jui-Ting Chien, Hwann-Tzong Chen, Tyng-Luh Liu
This paper presents a "learning to learn" approach to figure-ground image segmentation.
no code implementations • 25 Nov 2018 • Shou-Yao Roy Tseng, Hwann-Tzong Chen, Shao-Heng Tai, Tyng-Luh Liu
We present a generic and flexible module that encodes region proposals by both their intrinsic features and the extrinsic correlations to the others.
no code implementations • 14 Jul 2018 • Shou-Yao Roy Tseng, Hwann-Tzong Chen, Shao-Heng Tai, Tyng-Luh Liu
We introduce the concept of Non-Local RoI (NL-RoI) Block as a generic and flexible module that can be seamlessly adapted into different Mask R-CNN heads for various tasks.
no code implementations • CVPR 2018 • Hsien-Tzu Cheng, Chun-Hung Chao, Jin-Dong Dong, Hao-Kai Wen, Tyng-Luh Liu, Min Sun
Then, we concatenate all six faces while utilizing the connectivity between faces on the cube for image padding (i. e., Cube Padding) in convolution, pooling, convolutional LSTM layers.
no code implementations • CVPR 2018 • Hsien-Tzu Cheng, Chun-Hung Chao, Jin-Dong Dong, Hao-Kai Wen, Tyng-Luh Liu, Min Sun
Then, we concatenate all six faces while utilizing the connectivity between faces on the cube for image padding (i. e., Cube Padding) in convolution, pooling, convolutional LSTM layers.
no code implementations • 18 Jul 2017 • Tyng-Luh Liu
In this work, we propose a novel multi-view spectral clustering method for large-scale data.
no code implementations • 1 Dec 2015 • Tsung-Yu Lin, Tsung-Wei Ke, Tyng-Luh Liu
We address the problem of converting large-scale high-dimensional image data into binary codes so that approximate nearest-neighbor search over them can be efficiently performed.
no code implementations • 8 Apr 2015 • Jyh-Jing Hwang, Tyng-Luh Liu
We address the problem of contour detection via per-pixel classifications of edge point.
no code implementations • 22 Dec 2014 • Jyh-Jing Hwang, Tyng-Luh Liu
We address the problem of contour detection via per-pixel classifications of edge point.
no code implementations • NeurIPS 2008 • Yen-Yu Lin, Tyng-Luh Liu, Chiou-Shann Fuh
In solving complex visual learning tasks, adopting multiple descriptors to more precisely characterize the data has been a feasible way for improving performance.