Fangheng Guan

Fangheng Guan

Auto-generated profile, based on publicly available information

Fangheng Guan

College of Electronic Information and Automation, Civil Aviation University of China, Tianjin 300300, 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
17 June 2024 in Future Internet
Verified
Journal Article
Open access
PDF file
CFCheng Fang
Cheng Fang
FLFeifei Liang
Feifei Liang
TLTianchi Li
Tianchi Li
FGFangheng Guan
Fangheng Guan
The primary challenge in Multimodal sentiment analysis (MSA) lies in developing robust joint representations that can effectively learn mutual information from diverse modalities. Previous research in this field tends to rely on feature concatenation to obtain joint representations. However, these approaches fail to fully exploit interactive patterns to ensure consistency and differentiation across different modalities. To address this limitation, we propose a novel framework for multimodal sentiment analysis, named CDML (Consistency and Difference using a Multitask Learning network). Specifically, CDML uses an attention mechanism to assign the attention weights of each modality efficiently. Adversarial training is used to obtain consistent information between modalities. Finally, the difference among the modalities is acquired by the multitask learning framework. Experiments on two benchmark MSA datasets, CMU-MOSI and CMU-MOSEI, showcase that our proposed method outperforms the seven existing approaches by at least 1.3% for Acc-2 and 1.7% for F1.

5 September 2023 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
CFCheng Fang
Cheng Fang
MSMing Sun
Ming Sun
BHBo Huang
Bo Huang
FGFangheng Guan
Fangheng Guan
HZHonghao Zhou
Honghao Zhou
LYLei Yang
Lei Yang
The along-track resolution of conventional airborne synthetic aperture radar altimetry (SARAL), where the coherent processing interval (CPI) is the burst length, is not a full resolution due to beam limited. For a fully focused radar altimeter image, a novel airborne SARAL processing algorithm based on a long CPI is proposed in this paper. The long CPI is capable of achieving higher azimuthal resolution than that in the conventional altimeter with limited number of pulses. To conquer the problem of atmosphere turbulence and in accordance of motion deviation of the airborne platform, a sub-aperture phase gradient auto-focusing (PGA) framework is introduced to alleviate non-systematic phase errors (NsPE). In this framework, NsPE is estimated and compensated in along-track so that a fully-focused delayed Doppler map (DDM) can be guaranteed. Finally, the height parameter estimation is performed to the one-dimensional altimeter echoes after multi-looking processing of DDM to improve the estimation accuracy. Given that the conventional re-tracking algorithm is sensitive to noise, which may degrade the estimation accuracy, a flexible Bayesian method is designed in a hierarchical manner for the SARAL re-tracking. The SARAL raw data is utilized in the experiment. The results of image entropy and RMSE demonstrate the effectiveness and superiority of the proposed algorithm both qualitatively and quantitatively.