Haomin Yu

Dr. Haomin Yu


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
25 January 2025 in Entropy
Verified
Journal Article
Open access
PDF file
FLFengjiao Liang
Fengjiao Liang
QLQingyong Li
Qingyong Li
HYHaomin Yu
Haomin Yu
WWWen Wang
Wen Wang
Weakly supervised crack segmentation aims to create pixel-level crack masks with minimal human annotation, which often only differentiate between crack and normal no-crack patches. This task is crucial for assessing structural integrity and safety in real-world industrial applications, where manually labeling the location of cracks at the pixel level is both labor-intensive and impractical. Addressing the challenges of labeling uncertainty, this paper presents CrackCLIP, a novel approach that leverages language prompts to augment the semantic context and employs the Contrastive Language–Image Pre-Training (CLIP) model to enhance weakly supervised crack segmentation. Initially, a gradient-based class activation map is used to generate pixel-level coarse pseudo-labels from a trained crack patch classifier. The estimated coarse pseudo-labels are utilized to fine-tune additional linear adapters, which are integrated into the frozen image encoders of CLIP to adapt the CLIP model to the specialized task of crack segmentation. Moreover, specific textual prompts are crafted for crack characteristics, which are input into the frozen text encoder of CLIP to extract features encapsulating the semantic essence of the cracks. The final crack segmentation is determined by comparing the similarity between text prompt features and visual patch token features. Comparative experiments on the Crack500, CFD, and DeepCrack datasets demonstrate that the proposed framework outperforms existing weakly supervised crack segmentation methods, and the pre-trained vision-language model exhibits strong potential for crack feature learning, thereby enhancing the overall performance and generalization capabilities of the proposed framework.

1 September 2025
Verified
Conference Paper
Open access
PDF file
HWHaichen Wang
Haichen Wang
LYLiu Yang
Liu Yang
XZXinyuan Zhang
Xinyuan Zhang
HYHaomin Yu
Haomin Yu
MLMing Li
Ming Li
Jilin HuJilin Hu
Jilin Hu
Passenger demand forecasting helps optimize vehicle scheduling, thereby improving urban efficiency. Recently, attention-based methods have been used to adequately capture the dynamic nature of spatio-temporal data. However, existing methods that rely on heuristic masking strategies cannot fully adapt to the complex spatio-temporal correlations, hindering the model from focusing on the right context. These works also overlook the high-level correlations that exist in the real world. Effectively integrating these high-level correlations with the original correlations is crucial. To fill this gap, we propose the Aggregation Differential Transformer (ADFormer), which offers new insights to demand forecasting promotion. Specifically, we utilize Differential Attention to capture the original spatial correlations and achieve attention denoising. Meanwhile, we design distinct aggregation strategies based on the nature of space and time. Then, the original correlations are unified with the high-level correlations, enabling the model to capture holistic spatio-temporal relations. Experiments conducted on taxi and bike datasets confirm the effectiveness and efficiency of our model, demonstrating its practical value. The code is available at https://github.com/decisionintelligence/ADFormer.

22 August 2024
Verified
Book Chapter
Closed access
PDF file
HYHaomin Yu
Haomin Yu
YSYanru Song
Yanru Song
Jilin HuJilin Hu
Jilin Hu
CGChenjuan Guo
Chenjuan Guo
BYBin Yang
Bin Yang
CJChristian S. Jensen
Christian S. Jensen

8 March 2024 in The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Verified
Journal Article
Open access
PDF file
YCYuan Cheng
Yuan Cheng
JXJ. Xue
J. Xue
HYHaomin Yu
Haomin Yu
GHG. Hai
G. Hai
The Antarctic Ice Sheet (AIS) holds the potential to raise global sea levels by a staggering 57 m. Ice shelves, extensions of these ice sheets into the ocean, play a crucial role in stabilizing the upstream ice sheet. The Brunt Ice Shelf (BIS), located in the Southern Weddell Sea region and proximal to the continental shelf break, stands as a vital indicator for the region's response to global climate warming. Hence, this study systematically employs a range of remote sensing and field data, including remote sensing data from 1963 to 2023 (ARGON images, Landsat1-9 images, Sentinel-1 images, ICESat-2 altimetry data, etc.) and in-situ ice flow velocity data since 1956, to conduct a comprehensive analysis of the BIS's calving cycle and evolution over nearly six decades. The results reveal a significant exacerbation of both dynamical and structural instabilities in the BIS compared to the calving cycle in the 1970s. Combining analyses of air temperature from meteorological station, water temperature from ocean models and in-situ measurements such as Conductivity-Temperature-Depth profiles, the study suggests that the increased instability of the BIS is likely linked to the intrusion of warmer currents over the continental shelf. This research provides valuable insights into the evolving behavior of the BIS in response to changing climate conditions, underscoring the critical role of remote sensing data in advancing our understanding of these processes on a global scale.

9 June 2023 in arXiv
Verified
Preprint
Open access
PDF file
HYHaomin Yu
Haomin Yu
YSYanru Song
Yanru Song
Jilin HuJilin Hu
Jilin Hu
CGChenjuan Guo
Chenjuan Guo
BYBin Yang
Bin Yang
Crystal property prediction is a crucial aspect of developing novel materials. However, there are two technical challenges to be addressed for speeding up the investigation of crystals. First, labeling crystal properties is intrinsically difficult due to the high cost and time involved in physical simulations or lab experiments. Second, crystals adhere to a specific quantum chemical principle known as periodic invariance, which is often not captured by existing machine learning methods. To overcome these challenges, we propose the crystal-specific pre-training framework for learning crystal representations with self-supervision. The framework designs a mutex mask strategy for enhancing representation learning so as to alleviate the limited labels available for crystal property prediction. Moreover, we take into account the specific periodic invariance in crystal structures by developing a periodic invariance multi-graph module and periodic attribute learning within our framework. This framework has been tested on eight different tasks. The experimental results on these tasks show that the framework achieves promising prediction performance and is able to outperform recent strong baselines.

7 May 2025 in Environmental and Ecological Statistics
Verified
Journal Article
Open access
PDF file
HYHaomin Yu
Haomin Yu
CDChristian Damgaard
Christian Damgaard
To develop a predictive ecosystems model in a changing environment, it is important to require causal knowledge of the most important ecosystem process. More specifically, it is important to identify the causal environmental factors that drive vegetation changes in natural and semi-natural habitats managing and preserving ecosystems. In this study, we investigate spatio-temporal co-occurrence patterns between temporal changes in plant abundance data and possible causal environmental factors across multiple sites in wet heathlands. We explore two novel similarity measurement methods for temporal trends and environmental factors for assessing the potential influence of spatial proximity and specific environmental variables. Our results reveal several findings. First, close geographic proximity of wet heathland sites shows little association with vegetation change, suggesting the need to explore other influencing factors. Second, we observe a potential association between temporal correlation and environmental factors in specific instances, highlighting the need for further investigation. Furthermore, we identify significant Granger causality relationships at specific sites, highlighting overall productivity measured by NDVI as a potential causal factor for driving species vegetation changes in wet heathlands. Lastly, we identify a pattern where observation sites with high temporal and environmental similarity are often located near the coastline, suggesting potential maritime influences. These results underscore the complexity of ecosystem–environment interactions and emphasize the importance of analyzing comprehensive spatio-temporal environmental and ecological data when investigating causality.