no code implementations • 9 Sep 2024 • Fan Zhang, Lingling Li, Licheng Jiao, Xu Liu, Fang Liu, Shuyuan Yang, Biao Hou
In a series of FPN experiments on the scale-preferred tasks, we found that the ``divide-and-conquer'' idea of FPN severely hampers the detector's learning in the right direction due to the large number of large-scale negative samples and interference from background noise.
1 code implementation • 4 Aug 2024 • Zhihao LI, Biao Hou, Siteng Ma, Zitong Wu, Xianpeng Guo, Bo Ren, Licheng Jiao
We design a \textit{scaling center crop} operation to create the rotated crop with random orientation on each original image, introducing the explicit angle variation.
1 code implementation • 8 Jan 2022 • Helei Qiu, Biao Hou, Bo Ren, Xiaohua Zhang
And then a spatio-temporal tuples self-attention module is proposed to capture the relationship of different joints in consecutive frames.
2 code implementations • 21 Nov 2021 • Zhonghua Li, Biao Hou, Zitong Wu, Licheng Jiao, Bo Ren, Chen Yang
We convert a lightweight FCOSR model to TensorRT format, which achieves 73. 93 mAP on DOTA1. 0 at a speed of 10. 68 FPS on Jetson Xavier NX with single scale.
no code implementations • 7 May 2021 • Shuang Wang, Dong Zhao, Yi Li, Chi Zhang, Yuwei Guo, Qi Zang, Biao Hou, Licheng Jiao
Feature alignment between domains is one of the mainstream methods for Unsupervised Domain Adaptation (UDA) semantic segmentation.
no code implementations • IEEE Transactions on Cybernetics 2021 • Xu Liu, Lingling Li, Fang Liu, Biao Hou, Shuyuan Yang, Licheng Jiao
Second, the group spatial attention and group spectral attention modules are proposed to extract image features.
no code implementations • IEEE Transactions on Neural Networks and Learning Systems 2021 • Licheng Jiao, Ruohan Zhang, Fang Liu, Shuyuan Yang, Biao Hou, Lingling Li, Xu Tang
Video object detection, a basic task in the computer vision field, is rapidly evolving and widely used.
no code implementations • 5 Sep 2018 • Yan Ju, Lingling Li, Licheng Jiao, Zhongle Ren, Biao Hou, Shuyuan Yang
Due to the limited amount and imbalanced classes of labeled training data, the conventional supervised learning can not ensure the discrimination of the learned feature for hyperspectral image (HSI) classification.
no code implementations • 18 Jul 2017 • Zaidao Wen, Biao Hou, Qian Wu, Licheng Jiao
This paper develops a novel iterative framework for subspace clustering in a learned discriminative feature domain.
no code implementations • 30 Apr 2017 • Zaidao Wen, Biao Hou, Licheng Jiao
Linear synthesis model based dictionary learning framework has achieved remarkable performances in image classification in the last decade.
no code implementations • 24 Apr 2017 • Biao Hou, Zaidao Wen, Licheng Jiao, Qian Wu
Sparsity-regularized synthetic aperture radar (SAR) imaging framework has shown its remarkable performance to generate a feature enhanced high resolution image, in which a sparsity-inducing regularizer is involved by exploiting the sparsity priors of some visual features in the underlying image.