Da Ai

Da Ai

Auto-generated profile, based on publicly available information

Da Ai

Xian Key Laboratory of Image Processing Technology and Applications for Public Security, Xian University of Posts and Telecommunications, Xian, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
6 January 2025 in Expert Systems with Applications
Verified
Journal Article
Closed access
DADa Ai
Da Ai
SQSiyu Qin
Siyu Qin
ZNZihe Nie
Zihe Nie
DWDianwei Wang
Dianwei Wang
Hui YuanHui Yuan
Hui Yuan
YLYing Liu
Ying Liu

24 February 2025 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
DADa Ai
Da Ai
JWJiahao Wang
Jiahao Wang
THTing He
Ting He
Hui YuanHui Yuan
Hui Yuan
YLYing Liu
Ying Liu
NLNam Ling
Nam Ling
Introducing saliency information to mitigate perceptual redundancy and achieve superior compression represents a novel approach to the development of video compression. Existing saliency-based compression coding methods rely on the determination of saliency regions and focus too much on saliency regions while ignoring the perceptible distortion in non-saliency regions. We propose a spatiotemporal visual perceptual rate-distortion optimization (PRDO) algorithm for Versatile Video Coding (H.266/VVC) that is more in line with the human visual system (HVS). Firstly, we establish a linear weighted distortion model based on spatiotemporal and saliency features. The distortion model makes effective use of saliency features while considering image content in non-saliency regions that is still perceptible to the human eye, thereby achieving an overall visual effect that conforms to human subjective perception. Based on this distortion model, we propose a saliency adaptive quantization parameter (SAQP) selection method with a more flexible quantization parameter selection range, adaptively allocating the optimal coding unit quantization parameter according to the saliency regions of the image, ensuring a balanced bitrate allocation between saliency and non-saliency regions. The proposed method is implemented for the first time on the H.266/VVC coding standard, attaining an average bitrate saving of 19.9% across all test sequences and an average PSNR improvement of 2.65 dB in saliency regions compared to VTM16.0. The BD-EWPSNR of the proposed PRDO and SAQP method improves by 1.34 dB and 1.45 dB in the All-Intra and Lowdelay_P encoding modes, respectively. Additionally, the BD-Rate based on EWPSNR is reduced by 25.86% and 33.73%, respectively, with an overall compression coding time saving of 19.76%. The experimental results demonstrate that the proposed method can significantly reduce the bit rate and coding time while improving the subjective perceptive quality, providing a competitive solution for video compression coding.

1 January 2026 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
CSChang Sun
Chang Sun
Hui YuanHui Yuan
Hui Yuan
SJShiqi Jiang
Shiqi Jiang
DADa Ai
Da Ai
WZWei Zhang
Wei Zhang
RHRaouf Hamzaoui
Raouf Hamzaoui

27 April 2026 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
DADa Ai
Da Ai
TLTianzhen Liu
Tianzhen Liu
CHChenyang Han
Chenyang Han
DWDianwei Wang
Dianwei Wang
This paper proposes a recognition-oriented rate-distortion optimization (RO-RDO) method. As a novel Video Coding for Machines (VCM) approach, it addresses the challenge of degraded recognition performance for small targets, such as license plates, caused by coding distortion under stringent bitrate constraints, particularly in Unmanned Aerial Vehicle (UAV) applications. To support this research, we constructed the first Distance Calibrated License Plate Dataset (DCLP) in the field of VCM, establishing a solid foundation for future studies investigating the correlation between shooting distance, bitrate, and recognition performance. Experimental results demonstrate that, compared with the H.266/VVC standard, RO-RDO achieves 43.99% bitrate savings and reduces encoding time by 73.76% while maintaining comparable mean Average Precision (mAP). Under the same bitrate, the license plate recognition rate improves by 4.13%, and the mAP increases by 7.17%, with a 2.43 dB PSNR gain in the license plate region, effectively mitigating the impact of encoding distortion on the recognition rate.

1 January 2026 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
DWDianwei Wang
Dianwei Wang
KYKang Yang
Kang Yang
JFJie Fang
Jie Fang
YLYuanqing Li
Yuanqing Li
DADa Ai
Da Ai
JTJianing Tang
Jianing Tang

22 September 2023
Verified
Conference Paper
Open access
DADa Ai
Da Ai
XZXiaoyang Zhang
Xiaoyang Zhang
SQSiyu Qin
Siyu Qin
CXCe Xu
Ce Xu
YLYing Liu
Ying Liu
Point cloud semantic segmentation is widely used in the fields of autonomous driving and robotics. In order to achieve more effective semantic segmentation, a point cloud semantic segmentation network based on self-learning aggregation sampling and dual-feature attentive pooling (SADF-Net) is proposed. A self-learning aggregation sampling module is first designed so that the sampling points are no longer limited to a subset of the original point cloud and more feature information is retained. Then the contextual features between local regions are fully learned by a dual-feature attention pooling module. Finally, the semantic segmentation results are obtained by linear interpolation upsampling. The experimental results on the S3DIS dataset show that the mIoU of SADF-Net is improved by 27.2% compared with the backbone network PointNet, and the mloU is improved by 7.2% compared with the KVGCN network in recent years.