-
Domain adaptive pose estimation via multi-level alignment
Authors:
Yugan Chen,
Lin Zhao,
Yalong Xu,
Honglei Zu,
Xiaoqi An,
Guangyu Li
Abstract:
Domain adaptive pose estimation aims to enable deep models trained on source domain (synthesized) datasets produce similar results on the target domain (real-world) datasets. The existing methods have made significant progress by conducting image-level or feature-level alignment. However, only aligning at a single level is not sufficient to fully bridge the domain gap and achieve excellent domain…
▽ More
Domain adaptive pose estimation aims to enable deep models trained on source domain (synthesized) datasets produce similar results on the target domain (real-world) datasets. The existing methods have made significant progress by conducting image-level or feature-level alignment. However, only aligning at a single level is not sufficient to fully bridge the domain gap and achieve excellent domain adaptive results. In this paper, we propose a multi-level domain adaptation aproach, which aligns different domains at the image, feature, and pose levels. Specifically, we first utilize image style transer to ensure that images from the source and target domains have a similar distribution. Subsequently, at the feature level, we employ adversarial training to make the features from the source and target domains preserve domain-invariant characeristics as much as possible. Finally, at the pose level, a self-supervised approach is utilized to enable the model to learn diverse knowledge, implicitly addressing the domain gap. Experimental results demonstrate that significant imrovement can be achieved by the proposed multi-level alignment method in pose estimation, which outperforms previous state-of-the-art in human pose by up to 2.4% and animal pose estimation by up to 3.1% for dogs and 1.4% for sheep.
△ Less
Submitted 25 April, 2024; v1 submitted 23 April, 2024;
originally announced April 2024.
-
An end-to-end CNN framework for polarimetric vision tasks based on polarization-parameter-constructing network
Authors:
Yong Wang,
Qi Liu,
Hongyu Zu,
Xiao Liu,
Ruichao Xie,
Feng Wang
Abstract:
Pixel-wise operations between polarimetric images are important for processing polarization information. For the lack of such operations, the polarization information cannot be fully utilized in convolutional neural network(CNN). In this paper, a novel end-to-end CNN framework for polarization vision tasks is proposed, which enables the networks to take full advantage of polarimetric images. The f…
▽ More
Pixel-wise operations between polarimetric images are important for processing polarization information. For the lack of such operations, the polarization information cannot be fully utilized in convolutional neural network(CNN). In this paper, a novel end-to-end CNN framework for polarization vision tasks is proposed, which enables the networks to take full advantage of polarimetric images. The framework consists of two sub-networks: a polarization-parameter-constructing network (PPCN) and a task network. PPCN implements pixel-wise operations between images in the CNN form with 1x1 convolution kernels. It takes raw polarimetric images as input, and outputs polarization-parametric images to task network so as to complete a vison task. By training together, the PPCN can learn to provide the most suitable polarization-parametric images for the task network and the dataset. Taking faster R-CNN as task network, the experimental results show that compared with existing methods, the proposed framework achieves much higher mean-average-precision (mAP) in object detection task
△ Less
Submitted 18 April, 2020;
originally announced April 2020.
-
A quantitative evaluation of health care system in US, China, and Sweden
Authors:
Qixin Wang,
Menghui Li,
Hualong Zu,
Mingyi Gao,
Chenghua Cao,
Li Charlie Xia
Abstract:
This study is mainly aimed at evaluating the effectiveness of current health care systems of several representative countries and improving that of the US. To achieve these goals, a people-oriented non-linear evaluation model is designed. It comprises one major evaluation metric and four minor metrics. The major metric is constituted by combining possible factors that most significantly determine…
▽ More
This study is mainly aimed at evaluating the effectiveness of current health care systems of several representative countries and improving that of the US. To achieve these goals, a people-oriented non-linear evaluation model is designed. It comprises one major evaluation metric and four minor metrics. The major metric is constituted by combining possible factors that most significantly determine or affect the life expectancy of people in this country. The four minor metrics evaluate less important aspects of health care systems and are subordinate to the major one. The authors rank some of the health care systems in the world according to the major metric and detect problems in them with the help of minor ones. It is concluded that the health care system of Sweden scores higher than the US and Chinese system scores lower than that of the US. Especially, the health care system of US lags behind a little bit compared with its economic power. At last, it is reasonable for the American government to optimize the arrangement of funding base on the result of goal programming model.
△ Less
Submitted 19 April, 2013;
originally announced April 2013.
