HLHao Liu

Hao Liu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Hao Liu has a total of 52 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Diverse Style Prompt Assistant for Generalized Remote Sensing Object Classification

    by Diedie Jia, Wenda Zhao, Hao Liu, Haipeng Wang

    Due to the existence of domain shift between different remote sensing datasets, remote sensing object classification models trained on the source domain often face the challenge of reduced accuracy when applied to unseen domains. In this paper, we propose a diverse style prompt assistant for generalized remote sensing object classification. We utilize the powerful knowledge reserve of the CLIP model to solve the domain generalization problem from the perspective of text prompts. CLIP’s text features simulate feature-space style diversity, and then diverse styles guide the classification model to generate general image features. Specifically, we first implement a style diversity loss and maximum entropy loss to train prompts to obtain multiple style prompts. Then, we introduce the style prompts to guide the optimization process of image features cross-modally within the joint feature space, enabling the classification model to generate domain-invariant features. We conduct extensive experiments on three remote sensing datasets, and the results show that our method achieves superior generalization performance compared with the state of the arts. Codes are available at https://github.com/Jia-dd/DSPA.

  • AdaptiveCore: Adaptive Parallel Core Decomposition Framework

    by Chen Zhao, Zhigao Zheng, Hao Huang, Hao Liu, Dacheng Tao

    Core decomposition is a widely used hierarchical analysis algorithm for large-scale graphs. It achieves this decomposition by iteratively peeling the vertices along with their adjacency edges off into different hierarchies. With the timeliness requirements of modern applications, many researchers have introduced accelerators, particularly GPUs, to improve the computational efficiency of graph algorithms. However, the empty, sparse, and numerous hierarchies in large graphs lead to inefficient computation and parallelism, not only including unnecessary searching for the hierarchy’s vertices, but also significant thread wastage when peeling off the adjacency edges of these vertices. In this paper, we propose an adaptive parallel framework for core decomposition, named AdaptiveCore. First, it improves vertex searching efficiency by adaptively skipping the empty hierarchies and reducing the search space. Moreover, it greatly improves thread utilization by adaptively allocating the available threads to peel off the adjacency edges. Comprehensive experiments show that, compared with the state-of-the-art works, the proposed framework achieves an average speedup of $7.1\times$ on the GPU platform and up to $2.0\times$ on the multi-core CPU platform.

  • Lightweightability Measurement and a General Lightweight Design Framework for On-Orbit Image Interpretation Neural Networks

    by Yanhua Pang, Guoxu Zhou, Xin Meng, Haipeng Wang, Hao Liu, Xinlong Pan, Bo Chen

    Satellite on-orbit remote sensing image intelligent interpretation relies on on-orbit devices with extremely limited computational resources, utilizing advanced neural networks designed through lightweight methodologies to achieve fast and accurate interpretation of on-orbit remote sensing images. Nevertheless, the current design of lightweight neural networks exhibits three salient issues: firstly, the prevailing “one-size-fits-all" network lightweight design pattern is notably inefficient; secondly, there is a deficiency in analysing the impact of lightweight operations on network performance; thirdly, the mutual influence among various lightweight operations is overlooked. To address these issues, firstly, we propose a neural network lightweightability measurement model and its computational method by investigating the effects of various lightweight operations; secondly, we propose a neural network general lightweight design framework (GLD) tailored for satellite on-orbit remote sensing images intelligent interpretation. Specifically, GLD, based on a meta-leaning approach, integrates knowledge distillation (KD), pruning and quantization, three general lightweight technologies, into a framework. It dynamically assesses the distillability, prunability and quantifiability of neural networks, and uses this assessment and uses this as supervision to dynamically jointly optimizes KD, pruning and quantization, making it applicable to various mainstream neural networks; furthermore, we explore the mutual influence among lightweight operations based on GLD; finally, through ablation experiments and comparative experiments, we further verify the effectiveness and superiority of GLD.


  • Open Access

    Multimodal Semantic Segmentation in Yangtze River Economic Belt: Outcome of the 2024 IEEE WHISPERS MMSeg-YREB Challenge

    by Jiepan Li, He Huang, Yu Xia, Xian Xu, Yingxin Wu, Qi Guo, Yinhe Liu, Yanfei Zhong, Jeongho Min, Sungbin Son, et al.

    With the growing availability of remote sensing (RS) data from diverse platforms, multimodal RS techniques have emerged as a transformative solution for large-scale semantic segmentation. In response, we developed MMSeg-YREB, a specialized framework that integrates complementary RS modalities, such as multispectral and synthetic aperture radar (SAR) data from Sentinel-1/2 sources, to enhance the accuracy and robustness of land use and land cover (LULC) mapping across urban and regional landscapes within the Yangtze River Economic Belt (YREB). By leveraging extensive geographic coverage and heterogeneous data sources, MMSeg-YREB supports a wide range of applications, from precise urban planning to comprehensive environmental monitoring. Utilizing state-of-the-art artificial intelligence (AI) methodologies, this framework aims to develop highly generalizable and scalable semantic segmentation models, driving methodological advancements and accelerating the adoption of Earth observation (EO) technologies across diverse regions. As part of this initiative, the Multimodal Semantic Segmentation challenge, i.e., MMSeg-YREB, is organized in conjunction with the 14th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS) 2024. To foster further research and innovation, all datasets and code will be publicly released at https://github.com/danfenghong/Outcome-of-the-2024-IEEE-WHISPERS-MMSeg-YREB-Challenge for the sake of reproducibility, contributing to the broader EO and RS communities.


  • Open Access

    Risk Assessment of UAS Events From the Perspective of Heterogeneous Graphs

    by Yu An, Qingpei Yang, Hao Liu, Liang Geng

    With the widespread application of Unmanned Aerial Systems (UAS) in the aviation sector, accidents caused by drones have become increasingly frequent. As a result, numerous studies have been conducted by scholars on the risk assessment of UAS accidents. To address the challenge of evaluating the accident risks arising from the ever-changing operational environments and diverse tasks of UAS, this paper proposes a risk prediction framework for UAS accidents based on Heterogeneous Graph Embedding (HGE). By integrating various node attributes and the relationships between different factors, the framework constructs a heterogeneous graph from UAS aviation incident data. The complex heterogeneous relationships are transformed into homogeneous subgraphs through different path patterns, which are used to represent the data. Random walks are performed on each homogeneous subgraph to capture semantic information, and the Skip-Gram method is employed to embed the semantic attention-weighted walk results into high-dimensional vectors. These vectors are then integrated with the original graph embedding vectors to update the representations of incident nodes, enabling accurate prediction of accident severity and risk types. Furthermore, a meta-path-based filter module is designed to evaluate the key factors contributing to UAS accident risk by comparing the contributions of different meta-paths under various path patterns. Experimental results on a UAS incident dataset demonstrate that the proposed method effectively predicts UAS operational risks in real time, achieving an accuracy of 97.56% for accident severity prediction. Additionally, the method provides stable predictions when dealing with complex heterogeneous data. From the perspective of a novel graph structure, this research offers a new approach for the safety management of future UAS systems.


  • Top Cited Papers

    Efficient On-Orbit Remote Sensing Imagery Processing via Satellite Edge Computing Resource Scheduling Optimization

    by Qiangqiang Jiang, Lujie Zheng, Yu Zhou, Hao Liu, Qinglei Kong, Yamin Zhang, Bo Chen

    With the enormous scale of remote sensing imagery generation, on-orbit computing has become a crucial paradigm to enable near real-time processing. Due to the limited onboard resources and on-orbit power supply, satellite edge computing (SEC) is developed for satellite-ground collaboration, aiding on-orbit computation. However, the intermittent satellite-to-ground transmission link poses an efficiency challenge when collaborating SEC resources. Therefore, this paper proposes a Satellite Edge Computing resource scheduling technique for On-orbit Remote Sensing imagery processing (SECORS). First, we design a remote sensing mission-specific SEC architecture, which involves an offline-online satellite working mode. Subsequently, a computational resource scheduling model (SEC-RSM) is established, including the directed acyclic graph (DAG) model and mathematical problem formulation. Next, to obtain effective scheduling solutions, we develop an end-to-end algorithm leveraging the multi-agent proximal policy optimization and heuristic rule of the earliest finish time (SEC-MPH). Finally, we build a simulation SEC platform to carry out experiments and implement several methods as the comparison including multi-objective evolutionary algorithms, deep reinforcement learning approaches, and the scheme without optimization (baseline). Simulation results show that SECORS achieves 68.87% and 66.60% reductions in time and energy for on-orbit computation. Moreover, our method improves the energy efficiency ratio by three times and achieves high processing capacity with 548 pixels per unit of power (W) and time (ms).


  • Lock-Free Triangle Counting on GPU

    by Zhigao Zheng, Guojia Wan, Jiawei Jiang, Chuang Hu, Hao Liu, Shahid Mumtaz, Bo Du

    Finding the triangles of large scale graphs is a fundamental graph mining task in many applications, such as motif detection, microscopic evolution, and link prediction. The recent works on triangle counting can be classified into merge-based or binary search-based paradigms. The merge-based triangle counting paradigm locates the triangles using the set intersection operation, which suffers from the random memory access problem. The binary search-based triangle counting paradigm sets the neighbors of the source vertex of an edge as the lookup array and searches the neighbors of the destination vertex. There are lots of expensive lock operations needed in the binary search-based paradigm, which leads to low thread efficiency. In this paper, we aim to improve the triangle counting efficiency on GPU by designing a lock-free policy named Skiff to implement a hash-based triangle counting algorithm. In Skiff, we first design a hash trie data layout to meet the coalesced memory access model and then propose a lock-free policy to reduce the conflicts of the hash trie. In addition, we use a level array to manage the index of the hash trie to make sure the nodes of the hash trie can be quickly located. Furthermore, we implement a CTA thread organization model to reduce the load imbalance of the real-world graphs. We conducted extensive experiments on NVIDIA GPUs to show the performance of Skiff. The results show that Skiff can achieve a good system performance improvement than the state-of-the-art (SOTA) works.