Search

Your search keyword '"Yi, Zhonghua"' showing total 16 results
Start Over Author "Yi, Zhonghua"
16 results on '"Yi, Zhonghua"'

1. Real-World Computational Aberration Correction via Quantized Domain-Mixing Representation

Author

Jiang, Qi, Yi, Zhonghua, Gao, Shaohua, Gao, Yao, Qian, Xiaolong, Shi, Hao, Sun, Lei, Xu, Zhijie, Yang, Kailun, and Wang, Kaiwei

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Robotics, Electrical Engineering and Systems Science - Image and Video Processing, Physics - Optics
Abstract

Relying on paired synthetic data, existing learning-based Computational Aberration Correction (CAC) methods are confronted with the intricate and multifaceted synthetic-to-real domain gap, which leads to suboptimal performance in real-world applications. In this paper, in contrast to improving the simulation pipeline, we deliver a novel insight into real-world CAC from the perspective of Unsupervised Domain Adaptation (UDA). By incorporating readily accessible unpaired real-world data into training, we formalize the Domain Adaptive CAC (DACAC) task, and then introduce a comprehensive Real-world aberrated images (Realab) dataset to benchmark it. The setup task presents a formidable challenge due to the intricacy of understanding the target aberration domain. To this intent, we propose a novel Quntized Domain-Mixing Representation (QDMR) framework as a potent solution to the issue. QDMR adapts the CAC model to the target domain from three key aspects: (1) reconstructing aberrated images of both domains by a VQGAN to learn a Domain-Mixing Codebook (DMC) which characterizes the degradation-aware priors; (2) modulating the deep features in CAC model with DMC to transfer the target domain knowledge; and (3) leveraging the trained VQGAN to generate pseudo target aberrated images from the source ones for convincing target domain supervision. Extensive experiments on both synthetic and real-world benchmarks reveal that the models with QDMR consistently surpass the competitive methods in mitigating the synthetic-to-real gap, which produces visually pleasant real-world CAC results with fewer artifacts. Codes and datasets will be made publicly available., Comment: Codes and datasets will be made publicly available at https://github.com/zju-jiangqi/QDMR

Published
2024

3. FocusFlow: Boosting Key-Points Optical Flow Estimation for Autonomous Driving

Author

Yi, Zhonghua, Shi, Hao, Yang, Kailun, Jiang, Qi, Ye, Yaozu, Wang, Ze, Ni, Huajian, and Wang, Kaiwei

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Robotics, Electrical Engineering and Systems Science - Image and Video Processing
Abstract

Key-point-based scene understanding is fundamental for autonomous driving applications. At the same time, optical flow plays an important role in many vision tasks. However, due to the implicit bias of equal attention on all points, classic data-driven optical flow estimation methods yield less satisfactory performance on key points, limiting their implementations in key-point-critical safety-relevant scenarios. To address these issues, we introduce a points-based modeling method that requires the model to learn key-point-related priors explicitly. Based on the modeling method, we present FocusFlow, a framework consisting of 1) a mix loss function combined with a classic photometric loss function and our proposed Conditional Point Control Loss (CPCL) function for diverse point-wise supervision; 2) a conditioned controlling model which substitutes the conventional feature encoder by our proposed Condition Control Encoder (CCE). CCE incorporates a Frame Feature Encoder (FFE) that extracts features from frames, a Condition Feature Encoder (CFE) that learns to control the feature extraction behavior of FFE from input masks containing information of key points, and fusion modules that transfer the controlling information between FFE and CFE. Our FocusFlow framework shows outstanding performance with up to +44.5% precision improvement on various key points such as ORB, SIFT, and even learning-based SiLK, along with exceptional scalability for most existing data-driven optical flow methods like PWC-Net, RAFT, and FlowFormer. Notably, FocusFlow yields competitive or superior performances rivaling the original models on the whole frame. The source code will be available at https://github.com/ZhonghuaYi/FocusFlow_official., Comment: Accepted to IEEE Transactions on Intelligent Vehicles (T-IV). The source code of FocusFlow will be available at https://github.com/ZhonghuaYi/FocusFlow_official

Published
2023

4. Minimalist and High-Quality Panoramic Imaging with PSF-aware Transformers

Author

Jiang, Qi, Gao, Shaohua, Gao, Yao, Yang, Kailun, Yi, Zhonghua, Shi, Hao, Sun, Lei, and Wang, Kaiwei

Subjects
Computer Science - Computer Vision and Pattern Recognition, Electrical Engineering and Systems Science - Image and Video Processing, Physics - Optics
Abstract

High-quality panoramic images with a Field of View (FoV) of 360{\deg} are essential for contemporary panoramic computer vision tasks. However, conventional imaging systems come with sophisticated lens designs and heavy optical components. This disqualifies their usage in many mobile and wearable applications where thin and portable, minimalist imaging systems are desired. In this paper, we propose a Panoramic Computational Imaging Engine (PCIE) to achieve minimalist and high-quality panoramic imaging. With less than three spherical lenses, a Minimalist Panoramic Imaging Prototype (MPIP) is constructed based on the design of the Panoramic Annular Lens (PAL), but with low-quality imaging results due to aberrations and small image plane size. We propose two pipelines, i.e. Aberration Correction (AC) and Super-Resolution and Aberration Correction (SR&AC), to solve the image quality problems of MPIP, with imaging sensors of small and large pixel size, respectively. To leverage the prior information of the optical system, we propose a Point Spread Function (PSF) representation method to produce a PSF map as an additional modality. A PSF-aware Aberration-image Recovery Transformer (PART) is designed as a universal network for the two pipelines, in which the self-attention calculation and feature extraction are guided by the PSF map. We train PART on synthetic image pairs from simulation and put forward the PALHQ dataset to fill the gap of real-world high-quality PAL images for low-level vision. A comprehensive variety of experiments on synthetic and real-world benchmarks demonstrates the impressive imaging results of PCIE and the effectiveness of the PSF representation. We further deliver heuristic experimental findings for minimalist and high-quality panoramic imaging. Our dataset and code will be available at https://github.com/zju-jiangqi/PCIE-PART., Comment: Accepted to IEEE Transactions on Image Processing (TIP). The dataset and code will be available at https://github.com/zju-jiangqi/PCIE-PART

Published
2023

6. FocusFlow: Boosting Key-Points Optical Flow Estimation for Autonomous Driving

Author

Yi, Zhonghua, Shi, Hao, Yang, Kailun, Jiang, Qi, Ye, Yaozu, Wang, Ze, Ni, Huajian, and Wang, Kaiwei

Abstract

Key-point-based scene understanding is fundamental for autonomous driving applications. At the same time, optical flow plays an important role in many vision tasks. However, due to the implicit bias of equal attention on all points, classic data-driven optical flow estimation methods yield less satisfactory performance on key points, limiting their implementations in key-point-critical safety-relevant scenarios. To address these issues, we introduce a points-based modeling method that requires the model to learn key-point-related priors explicitly. Based on the modeling method, we present FocusFlow, a framework consisting of 1) a mix loss function combined with a classic photometric loss function and our proposed Conditional Point Control Loss (CPCL) function for diverse point-wise supervision; 2) a conditioned controlling model which substitutes the conventional feature encoder by our proposed Condition Control Encoder (CCE). CCE incorporates a Frame Feature Encoder (FFE) that extracts features from frames, a Condition Feature Encoder (CFE) that learns to control the feature extraction behavior of FFE from input masks containing information of key points, and fusion modules that transfer the controlling information between FFE and CFE. Our FocusFlow framework shows outstanding performance with up to ${+}44.5\%$ precision improvement on various key points such as ORB, SIFT, and even learning-based SiLK, along with exceptional scalability for most existing data-driven optical flow methods like PWC-Net, RAFT, and FlowFormer. Notably, FocusFlow yields competitive or superior performances rivaling the original models on the whole frame.

Published
2024
Full Text
View/download PDF

7. Effects of daidzein on growth performance, blood metabolites and meat quality of finishing Xianan beef cattle

Author

Huan Liang, Kehui Ouyang, Yi Zhonghua, Zuodong Chen, Jun Bai, Xiaozhen Song, Xu Lanjiao, Qu Mingren, Chanjuan Liu, Xiang H. Zhao, Ke Pan, and X. L. Qi

Subjects
chemistry.chemical_classification, endocrine system, 0303 health sciences, Antioxidant, Chemistry, medicine.medical_treatment, Daidzein, 0402 animal and dairy science, food and beverages, Fatty acid, 04 agricultural and veterinary sciences, Beef cattle, 040201 dairy & animal science, Feed conversion ratio, Transaminase, 03 medical and health sciences, chemistry.chemical_compound, Animal science, Genetics, medicine, Animal Science and Zoology, Dry matter, Intramuscular fat, Agronomy and Crop Science, 030304 developmental biology
Abstract

An experiment was conducted to determine the effects of supplementing different amounts of daidzein in a diet on the growth performance, blood biochemical parameters and meat quality of finishing beef cattle. Thirty finishing Xianan steers were distributed in three groups equilibrated by weight and fed three different dietary treatments (concentrate ratio = 80%): (1) control; (2) 500 mg/kg daidzein and (3) 1000 mg/kg daidzein, respectively. Steers were slaughtered after an 80-day feeding trial. Results showed that daidzein supplementation had no effect on the final body weight, average daily gain and feed conversion rate of steers. Steers fed with 1000 mg/kg daidzein had greater dry matter intake than those fed with control diets. Compared with the control group, the 1000 mg/kg daidzein group had a higher fat thickness, lower shear force and lightness. The pH, drip loss, cooking loss, redness (a*), yellowness (b*), moisture, ash, crude protein and intramuscular fat of the Longissimus dorsi muscle were unaffected by daidzein supplementation. Compared with the control group, the 1000 mg/kg daidzein group significantly increased the serum concentrations of insulin, free fatty acid and Glutamic-pyruvic transaminase. The 500 mg/kg daidzein group significantly increased the serum concentration of tetraiodothyronine compared with the control group. Supplemental daidzein did not affect the blood antioxidant ability and blood immune parameters in serum. In conclusion, daidzein supplementation above 500 mg/day modifies feed intake and metabolic and hormonal profile, with positive and negative effects on meat quality.

Published
2019

8. RNA‐Seq analysis reveals the potential molecular mechanisms of daidzein on adipogenesis in subcutaneous adipose tissue of finishing Xianan beef cattle

Author

Liang, Huan, primary, Xu, Lanjiao, additional, Zhao, Xianghui, additional, Pan, Ke, additional, Yi, Zhonghua, additional, Bai, Jun, additional, Qi, Xinglei, additional, Xin, Junping, additional, Li, Meifa, additional, Ouyang, Kehui, additional, Song, Xiaozhen, additional, Liu, Chanjuan, additional, and Qu, Mingren, additional

Published
2019
Full Text
View/download PDF

9. RNA-Seq analysis reveals the potential molecular mechanisms of daidzein on adipogenesis in subcutaneous adipose tissue of finishing Xianan beef cattle

Author

Kehui Ouyang, Xinglei Qi, Chanjuan Liu, Lanjiao Xu, Junping Xin, Xianghui Zhao, Xiaozhen Song, Meifa Li, Huan Liang, Mingren Qu, Ke Pan, Yi Zhonghua, and Jun Bai

Subjects
Male, endocrine system, 040301 veterinary sciences, Subcutaneous Fat, Beef cattle, 0403 veterinary science, chemistry.chemical_compound, Food Animals, Downregulation and upregulation, Adipocyte, Animals, Glycolysis, RNA-Seq, Chemistry, Daidzein, 0402 animal and dairy science, food and beverages, Lipid metabolism, 04 agricultural and veterinary sciences, 040201 dairy & animal science, Animal Feed, Isoflavones, Cell biology, Diet, Gluconeogenesis, Gene Expression Regulation, Adipogenesis, Body Composition, Animal Science and Zoology, Cattle
Abstract

Daidzein has been reported to be effective in regulating lipid metabolism in animals. However, the molecular mechanisms of daidzein on adipogenesis in beef cattle are not yet reported and the results of daidzein on affecting lipid metabolism in other species have been conflicting. High-throughput sequencing of mRNA (RNA-Seq) technology was performed to elucidate the underlying molecular mechanisms of daidzein on adipogenesis in subcutaneous adipose tissue of finishing Xianan beef cattle. A total of 893 differentially expressed genes (DEGs) were identified by differential expression analysis, among which 405 genes were upregulated and 488 genes were downregulated. Bioinformatics analysis suggested that these DEGs were significantly enriched to the pathways related to lipid metabolism including ECM-receptor interaction, Glycolysis/Gluconeogenesis and Hedgehog signalling pathway. Daidzein significantly affected the candidate genes (Shh, Pec, Gli, Wnt6, DLK, IGFBP2, ID3 and C/EBPE) related to adipocyte differentiation. Besides, daidzein improved the ability of subcutaneous adipocytes in synthesizing triglycerides by directly using the long-chain fatty acids and enhanced the efficiency of triglyceride synthesis of subcutaneous adipocytes in Xianan steers. In conclusion, daidzein plays a positive role not only in adipogenic differentiation, but also in triglyceride synthesis in subcutaneous adipose tissue of Xianan beef cattle.

Published
2019

10. Expressing Emotional Concept in Visual Communication Design: An Anthropological Case Study on Emotional Design in China

Author

Yi Zhonghua, Liu Yifeng, and Li Wei

Subjects
Emotional design, Anthropology, 05 social sciences, Applied psychology, 050301 education, Questionnaire, 0501 psychology and cognitive sciences, Visual communication, Psychology, China, 0503 education, Social psychology, 050107 human factors
Abstract

By adopting visual communication design as the medium for introducing emotional designs, this paper investigates the relationship between emotional concepts and visual communication design. Applica...

Published
2016

11. RNA‐Seq analysis reveals the potential molecular mechanisms of daidzein on adipogenesis in subcutaneous adipose tissue of finishing Xianan beef cattle.

Author

Liang, Huan, Xu, Lanjiao, Zhao, Xianghui, Pan, Ke, Yi, Zhonghua, Bai, Jun, Qi, Xinglei, Xin, Junping, Li, Meifa, Ouyang, Kehui, Song, Xiaozhen, Liu, Chanjuan, and Qu, Mingren

Subjects
BEEF cattle, LIPID metabolism, ADIPOSE tissues, FATTY acids, DAIDZEIN, FAT cells
Abstract

Daidzein has been reported to be effective in regulating lipid metabolism in animals. However, the molecular mechanisms of daidzein on adipogenesis in beef cattle are not yet reported and the results of daidzein on affecting lipid metabolism in other species have been conflicting. High‐throughput sequencing of mRNA (RNA‐Seq) technology was performed to elucidate the underlying molecular mechanisms of daidzein on adipogenesis in subcutaneous adipose tissue of finishing Xianan beef cattle. A total of 893 differentially expressed genes (DEGs) were identified by differential expression analysis, among which 405 genes were upregulated and 488 genes were downregulated. Bioinformatics analysis suggested that these DEGs were significantly enriched to the pathways related to lipid metabolism including ECM–receptor interaction, Glycolysis/Gluconeogenesis and Hedgehog signalling pathway. Daidzein significantly affected the candidate genes (Shh, Pec, Gli, Wnt6, DLK, IGFBP2, ID3 and C/EBPE) related to adipocyte differentiation. Besides, daidzein improved the ability of subcutaneous adipocytes in synthesizing triglycerides by directly using the long‐chain fatty acids and enhanced the efficiency of triglyceride synthesis of subcutaneous adipocytes in Xianan steers. In conclusion, daidzein plays a positive role not only in adipogenic differentiation, but also in triglyceride synthesis in subcutaneous adipose tissue of Xianan beef cattle. [ABSTRACT FROM AUTHOR]

Published
2020
Full Text
View/download PDF

12. Stability of Solid-Solution Phase and the Nature of Phase Separation in Ru–Zr–O Ternary Oxide

Author

Bo Wu, Yi Zhonghua, Tang Dian, Wei Lin, Zhongzhi Tang, Xin Wang, and Junqiu Zhu

Subjects
Materials science, Spinodal decomposition, Thermal decomposition, Oxide, Thermodynamics, Ruthenium oxide, Surfaces, Coatings and Films, Electronic, Optical and Magnetic Materials, Amorphous solid, chemistry.chemical_compound, General Energy, chemistry, Phase (matter), Physical and Theoretical Chemistry, Phase diagram, Solid solution
Abstract

Equilibrium phase diagram of ruthenium–zirconium oxide (Ru–Zr–O) was calculated by a combination of ab initio density functional theory and thermodynamic calculations. The phase diagrams suggest that the solubility between ruthenium oxide and zirconium oxide is very low under normal experimental conditions, in good agreement with the prior reports in literature. Also provided is the theoretical support for the reported phenomenon that doping or amorphization of Ru–Zr–O might suppress phase separation of the oxide. To study the spinodal decomposition of Ru–Zr–O, we successfully prepared an amorphous Ru0.48Zr0.52O2 film of a solid solution phase by a thermal decomposition at a low temperature (563 K). In situ transmission electron microscopy was utilized to witness the spinodal decomposition of the amorphous Ru0.48Zr0.52O2 solid solution by electron–beam annealing. The present fundamental study of the phase behavior of Ru–Zr–O provides a guideline for the phase and microstructure design of Ru-based mixed ox...

Published
2012

13. Effect of Chinese herbal additive on growth performance and tissue melanin content in Taihe silky fowls

Author

Pan Ke, Li GuanHong, You JinMing, Qu MingRen, and Yi ZhongHua

Subjects
Polygonum, Ecology, Traditional medicine, business.industry, Fowl, Diarrhea Frequency, food and beverages, Soil Science, Plant Science, Biology, biology.organism_classification, Biotechnology, Melanin, Ingredient, Crinis Carbonisatus, Sesamum, business, Medicinal plants, Agronomy and Crop Science, Ecology, Evolution, Behavior and Systematics
Abstract

The effects of three different Chinese herbal additives on growth performance and content of melanin in tissues of Taihe silky fowls were evaluated to determine the feasibility of improving medicinal values of the fowls. Birds were fed basal diet (control group), basal diet supplemented with 1% Chinese herbal additive Ⅰ with Sesamum indicum L. as main ingredient (treatment group Ⅰ), 1% Chinese herbal additive Ⅱ with Polygonum multiflorum Thunb as main ingredient (treatment group Ⅱ) and 1% Chinese herbal additive Ⅲ with Crinis carbonisatus as main ingredient (treatment groupⅢ), respectively. Results show no significant dif- ferences in average daily gain, daily feed intake, and feed to gain ratio (P>0.05) of Taihe silky fowls aged 0~12 weeks among the 4 treatment groups. With the exception of groupⅢ, mortality rate in Chinese herbal additive groups generally tends to decline in com- parison with the control group, and diarrhea frequency in all Chinese herbal additive treatment groups is lower than that in the control. The Chinese herbal additive treatment groups have higher melanin content in pectoral muscle, leg muscle and skin than the control as the Taihe silky fowl is 4 weeks old. At 8 and 12 weeks old, content of melanin in Chinese herbal additive treatment groups is gener- ally higher than in the control. However, skin melanin content is slightly lower in treatment group Ⅱ than in the control at 8 weeks old, and at 12 weeks old treatment groups Ⅰ and Ⅱ have lower melanin content in leg muscle than control group does. Three dif- ferent Chinese herbal additive supplements at 1% total diet do not affect growth performance in general, but increase (by different degrees) tissue melanin content of Taihe silky fowls.

Published
2009

16. Minimalist and High-Quality Panoramic Imaging With PSF-Aware Transformers.

Author

Jiang Q, Gao S, Gao Y, Yang K, Yi Z, Shi H, Sun L, and Wang K

Abstract

High-quality panoramic images with a Field of View (FoV) of 360° are essential for contemporary panoramic computer vision tasks. However, conventional imaging systems come with sophisticated lens designs and heavy optical components. This disqualifies their usage in many mobile and wearable applications where thin and portable, minimalist imaging systems are desired. In this paper, we propose a Panoramic Computational Imaging Engine (PCIE) to achieve minimalist and high-quality panoramic imaging. With less than three spherical lenses, a Minimalist Panoramic Imaging Prototype (MPIP) is constructed based on the design of the Panoramic Annular Lens (PAL), but with low-quality imaging results due to aberrations and small image plane size. We propose two pipelines, i.e. Aberration Correction (AC) and Super-Resolution and Aberration Correction (SR&AC), to solve the image quality problems of MPIP, with imaging sensors of small and large pixel size, respectively. To leverage the prior information of the optical system, we propose a Point Spread Function (PSF) representation method to produce a PSF map as an additional modality. A PSF-aware Aberration-image Recovery Transformer (PART) is designed as a universal network for the two pipelines, in which the self-attention calculation and feature extraction are guided by the PSF map. We train PART on synthetic image pairs from simulation and put forward the PALHQ dataset to fill the gap of real-world high-quality PAL images for low-level vision. A comprehensive variety of experiments on synthetic and real-world benchmarks demonstrates the impressive imaging results of PCIE and the effectiveness of the PSF representation. We further deliver heuristic experimental findings for minimalist and high-quality panoramic imaging, in terms of the choices of prototype and pipeline, network architecture, training strategies, and dataset construction. Our dataset and code will be available at https://github.com/zju-jiangqi/PCIE-PART.

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results