WLWangchenxiao Liu

Wangchenxiao 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


Wangchenxiao Liu has a total of 4 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Reliability-Aware Dual-Stream Self-Training for Semi-Supervised Semantic Segmentation of High-Resolution Remote Sensing Imagery

    by Wangchenxiao Liu, Yan Lin, Wei Liu, Zhengquan Chen

    Semantic segmentation of high-resolution remote sensing imagery remains constrained by the high cost of pixel-level annotation. Semi-supervised learning can reduce this dependence by exploiting unlabeled images, but improving pseudo-label reliability often comes at the cost of reduced data coverage. Selecting only stable samples can suppress noise but may exclude complex scenes and minority classes, whereas retaining the complete unlabeled set preserves data diversity but introduces less reliable and class-biased supervision. To address this trade-off, we propose RAPST, a reliability-aware dual-stream self-training framework that combines reliable offline supervision with full-data online learning. In the offline stream, the Class-Aware Image Stability Gate (ISG) selects prediction-stable images while preserving class coverage. In the online stream, EMA-Smoothed Class-Adaptive Pseudo-Label Thresholding (CPT) adapts pixel-selection thresholds according to class-wise pseudo-label statistics. Reliability-Aware Prototype-Guided Category Contrast (PCC) further integrates reliable feature–label pairs from both streams to improve feature discrimination. Experiments on the ISPRS Vaihingen, ISPRS Potsdam, and WHDLD datasets under four annotation ratios show that RAPST achieves the best performance in most dataset–annotation settings, with mIoU improvements of 0.33–6.08 percentage points over the supervised-only baseline. Ablation and mechanism analyses further show that prediction stability is associated with pseudo-label quality, CPT increases supervision coverage for difficult classes with limited precision loss, and PCC improves feature-space separability.

  • Open Access

    Explainable Vision Analytics for Adaptive Campus Design: Diagnosing Multi-Dimensional Perceptual Differences

    by Yan Lin, Wangchenxiao Liu, Xi Sun

    Campus streetscapes are a key part of universities’ everyday public realm, yet the same scene may be perceived positively in one dimension while negatively in another. To diagnose such multi-dimensional perceptual differences and translate them into actionable design evidence, this study develops an interpretable vision analytics framework for adaptive campus design. Using 72,733 Baidu Street View images collected from 41 campuses in mainland China, the study integrates ResNet-50-based perception prediction, spatial element extraction, XGBoost–SHAP-based mechanism interpretation, Kruskal–Wallis H testing, and GIS-based scene mapping. Supported by supplementary in situ validation, six types of multi-dimensional perceptual differences were identified. Sky, buildings, vegetation, hardscape, and terrain were found to be the five most important spatial elements overall, among which sky, buildings, and vegetation repeatedly emerged as the dominant core elements distinguishing different perceptual types. These elements do not act independently or linearly, but jointly shape different types of multi-dimensional perceptual differences through nonlinear threshold effects and interactions. These perceptual difference types were further found to cluster in recognizable campus scenes, including main roads, plazas, lawns, forest belts, and lakeside spaces. Based on these findings, scene-specific piecemeal optimization strategies were derived to support the coordinated enhancement of perceived safety, liveliness, and beauty. Overall, the study shows that campus perception is shaped by holistic spatial configurations rather than the simple accumulation of isolated elements, and provides a quantitative basis for iterative, feedback-oriented adaptive campus design.