Purpose: In three-dimensional (3D) gait analysis, foot positioning during static calibration posture critically influences the calculation of biomechanical variables during dynamic tasks. Although stance width is known to alter lower extremity alignment, it remains unclear whether variations in stance width during calibration affect subsequent lower extremity kinematics and kinetics during running. This study aimed to evaluate the impact of calibration stance width variations on calculated lower extremity mechanics during running.Methods: Twenty healthy participants (10 males and 10 females; age: 24.7 ± 1.3 years; height: 1.73 ± 0.08 m; weight: 66.5 ± 10.7 kg) performed three static calibration postures: (1) shoulder-width stance width (SW), (2) wider stance width (WSW), and (3) narrower stance width (NSW), prior to running at a self-selected speed. Repeated measures ANOVA with statistical parametric mapping (SPM) was used to analyze continuous time-series data for hip, knee, and ankle joint angles and moments.Results: Compared to SW condition, the NSW condition demonstrated significantly greater ankle dorsiflexion angles, inversion angles, and external rotation moments. Conversely, the WSW condition demonstrated reduced ankle dorsiflexion and inversion angles, decreased external rotation moments, and increased ankle inversion moments. Furthermore, the WSW condition yielded significantly greater hip abduction and external rotation moments compared to the SW condition.Conclusions: These results demonstrate that variations in static calibration stance width introduce systematic biases into measurement outcomes and potentially alter biomechanical interpretations. Therefore, it is important that future studies standardize calibration stance width to minimize measurement artifacts and ensure robust biomechanical assessments.

Although fatigue is widely recognized as a risk factor for hamstring strain injuries, the mechanisms by which fatigue influences these injuries remain unclear. It is also uncertain whether fatigue differentially affects the risk factors for hamstring strain between lower limbs. Nineteen male track-and-field athletes underwent a fatigue intervention, with optimal muscle length assessed pre- and post-fatigue. Key biomechanical parameters during sprinting were also analyzed before and after fatigue. Post-fatigue, the non-dominant limb showed significant reductions in optimal lengths across all three biarticular hamstring muscles (p < 0.01; Cohen’s d = 0.95–1.07), with greater changes than the dominant limb (p < 0.05; η_p^2 = 0.16–0.24). No significant differences were observed in maximal hip–knee angles, peak muscle lengths, or the timing of peak muscle strain (p > 0.05). However, the non-dominant limb exhibited increased peak strain in the long head of the biceps femoris and semimembranosus (p < 0.05; Cohen’s d = 0.46–0.52), but not in the semitendinosus (p = 0.06; Cohen’s d = 0.90). No significant inter-limb differences were observed in the strain changes of the hamstring biarticular muscles (p > 0.05). Fatigue was not associated with changes in sprint-phase peak lengths of the non-dominant biarticular hamstring muscles; however, it was accompanied by a reduction in optimal length, which may have increased peak strain. Clinically, monitoring fatigue-related changes in hamstring functional properties and addressing potential inter-limb asymmetries may be relevant for injury prevention and rehabilitation. The same fatigue protocol was not associated with increased peak strain on the dominant side, indicating a potential limb-specific susceptibility to fatigue-related strain increases.

Nitrogen oxide (NOx) necessitates low-concentration detection at elevated temperatures due to its toxicity for both human health and sensing materials. This work demonstrates the use of Ni-doped Co3O4 frameworks (NCO) as high-performance sensing materials, that are derived from Ni-ZIF-67 via thermal annealing. Among that, the NCO-2 sample, featuring a hierarchical hollow structure with a specific surface area of 94.89 m2/g, enables sensitive detection of low-concentration NOx, especially 500 ppb NO2 at 150 degrees C with a high response value of Ra/ Rg-1 = 5.35 and high selectivity (SNO2/SCO=26.75). Characterization techniques and DFT calculations combined to confirm that Ni2+ substitute for Co2+ in tetrahedral sites of the Co3O4 is a strong adsorption sites for NO2 molecules with a high adsorption enthalpy of-3.04 eV and + 0.44 e transfer. Ex situ XPS analysis further elucidates NO2 poisoning mechanisms by identifying the formation of stable NO3-/NO2-species on the sensor surface at high concentrations, and provided corresponding solutions. To bridge laboratory research and practical application, a portable NO2 detector was developed by integrating an NCO-2 sensor array with smartphone connectivity. This device enables real-time monitoring of environmental NO2 pollution. Collectively, this study presents an integrated system that advances NO2 sensor research by combining material design, mechanistic insight, and prototype development.

Micro-nano bubbles (MNBs) enhance the oxidative capacity of water by promoting hydroxyl radical (& centerdot;OH) generation. This study investigated the regulatory effects of the gas-liquid ratio (Qg/Ql), aeration rate (v), and pH on & centerdot;OH generation by air MNBs generated via a pressurized gas dissolution method, with further validation in sulfur-containing wastewater (SC-W). The results indicated that a lower gas-liquid ratio favored nanobubble (NB) formation and consequently enhanced the cumulative & centerdot;OH yield. Increasing the gas-liquid ratio from 1% to 5% reduced NB concentration by a factor of 6.5 and decreased the cumulative & centerdot;OH yield by a factor of 4.4. Varying the aeration rate from 13.3 m s-1 to 17.0 m s-1 resulted in a non-monotonic change in NB quantity, which first increased and then decreased. Weakly alkaline conditions were most favorable for cumulative & centerdot;OH yield via NB shrinkage and collapse, while strong acidic or alkaline conditions suppressed & centerdot;OH generation by 83-96%. The optimal cumulative & centerdot;OH yield of 1.18 mu mol L-1 was achieved under the conditions of a 1% gas-liquid ratio, an aeration rate of 15.5 m s-1, and pH 7.7. The enhancement mechanism for sulfide removal was attributed to a synergistic effect between & centerdot;OH oxidation and oxygen mass transfer, with the dominant mechanism shifting from & centerdot;OH oxidation to mass transfer enhancement as the sulfide concentration increased. Under optimal conditions, the S2- removal rate reached 86.53% within 1 hour for industrial SC-W. This study establishes a foundation for optimizing the oxidative capacity of MNBs and provides technical support for the low-carbon treatment of industrial SC-W.

Hydrogen sulfide (H2S) is a harmful substance existing widely, it exerts significant impacts on both ecosystems and human beings, requiring timely detection. A ZnFe2O4@ZnO (ZFO@ZnO) p-n heterojunction was synthesized by using MIL-88A(Fe) as a precursor. p-ZnFe2O4 and n-ZnO serve as the core and shell layer, respectively. The energy band structure is reconstructed, resulting in formation of an internal built-in electric field between p and n region, reducing the bandgap width further inducing a large number of oxygen vacancy at the interface of heterojunctions. The existence of the p-n heterojunction and the internal built-in electric field was directly confirmed by HRTEM image, I-V curve, and UV-Vis DRS data. Gas sensing tests revealed that this sample exhibits ultrahigh performance toward H2S, with a response of 58.9-1 ppm. It also features ultra-fast response and recovery characteristics, enabling real-time detection of low-concentration H2S (100 ppb). Ex-situ XPS and EPR of the sample treated with HAS was employed to reveal the H2S sensing mechanism. It is found a small amount of ZnS and S0 are formed during the sensing process, and these two substances can be removed by heating them in air, thereby achieving long-term H2S sensing stability. In addition, we verified the practical application value of this gas sensor in terms of food spoilage. After a continuous experiment, the device can well distinguish spoiled eggs and fresh eggs. In summary, this work has not only been thoroughly investigated at the experimental level, but also achieved the transition from laboratory to application.

Cobalt-based spinel oxides are emerging as low-cost and selective materials for gas sensing, though their performance is still unsatisfactory. In this study, we developed a refined method to produce (Cu0.2Co0.8)Co2O4 frameworks derived from Cu-doped ZIF-67. This unique synthesis method imparts a distinctive morphology to the (Cu0.2Co0.8)Co2O4, featuring a high density of surface oxygen vacancies and doped Cu2+ ions bring stronger interaction between H2S molecules and the (Cu0.2Co0.8)Co2O4 coating, resulting in exceptional H2S gas sensing performance, including a high response value (Delta R/Ra = 5.11-500 ppb H2S), rapid response-recovery times (tau res = 3.6 s, tau res = 10.8 s) and an ultra-low detection limit (50 ppb). These results demonstrate the promising potential of this material for H2S gas detection applications. Ex situ XPS characterization reveals the gas sensing reaction mechanism, while density functional theory calculations confirm that the presence of Cu2+ significantly reduces the H2S adsorption enthalpy, thereby enhancing the overall gas sensing performance. This work not only introduces an approach to provide a better H2S gas sensor, but also paves new pathways for advanced inorganic synthesis methods for metal oxide catalysts.

Early ship trajectory prediction improves traffic coordination but increases the risk of intent misjudgment at waterway intersections, leading to deviations between predicted and actual trajectories. To address this, we propose a ship trajectory prediction model grounded in the International Maritime Organization (IMO) framework and the rule of "intent report - ship maneuver - trajectory change" observed in real-world waterway intersections. Our method enables early intent recognition by leveraging intent information embedded in shipshore speech communication. Within a defined spatiotemporal range, we associate communication data with observed trajectories to identify reported intentions. The extracted intent labels are integrated with encoded historical trajectory features and fed into a decoder, dynamically constraining predicted directions. This alignment with reported intent advances the prediction timeline without compromising accuracy. Empirical validation at the Wusongkou Estuary (Shanghai, China) demonstrates that our model advances the prediction timeline by 6.94-8.4 min compared to existing models, while maintaining similar accuracy. This work pioneers the integration of ship-shore speech communication into trajectory prediction, highlighting the potential of AIdriven maritime safety systems.

Bladder cancer, predominantly urothelial carcinoma, is a global health issue with increasing incidences and mortality. It poses significant diagnostic and therapeutic challenges due to its molecular heterogeneity and the limitations of current detection methods. Extracellular vesicles (EVs), including exosomes, play a crucial role in intercellular communication and have emerged as potential biomarkers and therapeutic agents in bladder cancer. This review focuses on the multifaceted roles of EVs in bladder cancer biology, their potential as diagnostic biomarkers, and their use in therapeutic strategies. We discuss how EVs reflect molecular subtypes of bladder cancer, participate in metabolic reprogramming and angiogenesis, and modulate cellular behavior. The review also highlights the advances in proteomic analysis of urinary and tissue-exudative EVs, identifying specific proteins and RNAs that could serve as non-invasive diagnostic markers. Furthermore, we explore the innovative use of EVs as natural nanocarriers for drug delivery in bladder cancer treatment, demonstrating their potential to enhance the efficacy of chemotherapy and selectively target cancer cells. The integration of EV-based diagnostics with traditional methods could lead to more personalized and effective bladder cancer management, emphasizing the need for further research and clinical validation.

The role of adipose-derived stromal stem cells (ADSCs) in BLCA progression is unclear. We investigated the effects of invasion, stemness, Epithelial-mesenchymal transition (EMT), and drug resistance of BLCA cells co-cultured with ADSCs for a long period of time. Cells were divided into six groups: ADSCs group, ADSCs: T24 group (10:1, 3:1 and 1:1 groups), ADSCs-derived conditioned medium group (CM) and T24 cell group (T24), and cells in each group were cultured to 14 days, and puromycin (puro) was added to the co-cultured cell to remove ADSCs cells without puro resistance, and then the function of T24 cells (10:1-COC, 3:1-COC and 1:1-COC) after co-culture was studied; CCK-8 assay, Transwell, Wound healing, Flow cytometry, RNA-sequencing, qRT-PCR and Western Bloting assay were used to detect cell proliferation, invasion, migration, apoptosis and cellular mRNA and protein expression levels, respectively. We unexpectedly found enhanced stemness and drug resistance of BLCA cells after prolonged contact culture with ADSCs. T24 cells after co-culture mediated cell proliferation, invasion, EMT, stemness, drug resistance and immune escape by up-regulating MDM2, mt-P53 and PD-L1, compared to CM and T24 groups. The inhibitor Atezo and CP-31,398 eliminated mt-P53 and PD-L1-mediated T24 cell drug resistance and stemness, respectively. This study demonstrated that after prolonged co-culture of BLCA cells with ADSCs, the stemness, drug resistance, and immune evasion of T24 cells were dramatically enhanced, suggesting that long-term resident ADSCs in the bladder cancer tumor microenvironment play a procarcinogenic role.

SnO2-based semiconductor gas-sensing materials are regarded as some of the most crucial sensing materials, owing to their extremely high electron mobility, high sensitivity, and excellent stability. To bridge the gap between laboratory-scale SnO2 and its industrial applications, low-cost and high-efficiency requirements must be met. This implies the need for simple synthesis techniques, reduced energy consumption, and satisfactory gas-sensing performances. In this study, we utilized a surfactant-free simple method to modify SnO2 nanoparticles with PdPt noble metals, ensuring the stable state of the material. Under the synergistic catalytic effect of Pd and Pt, the composite material (1.0 wt%-PdPt-SnO2) significantly enhanced its response to HCHO. This modification decreased the optimal working temperature to as low as 180 °C to achieve a response value (Ra/Rg = 8.2) and showcased lower operating temperatures, higher sensitivity, and better selectivity to detect 10 ppm of HCHO when compared with pristine SnO2 or single noble metal-decorated SnO2 sensors. Stability tests verified that the gas sensor signals based on PdPt-SnO2 nanoparticles exhibit good reliability. Furthermore, a portable HCHO detector was designed for practical applications, such as in newly purchased cushions, indicating its potential for industrialization beyond the laboratory.

Shear Flow Deformability Cytometry (SDC), a subset of microfluidics-based deformability cytometry methods, has been widely applied for assessing cellular mechanical properties due to its high throughput. SDC characterizes cells' physical information, including size and stiffness, using automated image processing techniques. However, traditional and convolutional neural network (CNN)-based image processing methods face challenges in achieving rapid and accurate cell segmentation in SDC, primarily due to low image contrast and short exposure times. In response, this paper introduces a novel cell segmentation network based on a PolarMask algorithm. This network leverages contour representation and image enhancement to enhance cell saliency. By incorporating Intersection over Union (IoU) prediction and conducting joint training, the network significantly boosts segmentation accuracy while simplifying network structures to reduce computational resource demands. The proposed network achieves a precision level of 0.925, a remarkable improvement of 0.228 compared to the PolarMask's precision of 0.697. Notably, these accuracy gains are attained while maintaining a mean Frames Per Second (FPS) rate of 12.5.

In the rapidly evolving domain of software development, the security and reliability of the open source software supply chain are of increasing concern. Outdated dependencies pose significant risks to the software ecosystem. This study aims to quantitatively reveal the prevalence of outdated dependencies in the open source supply chain, evaluate their impact on software security, and identify effective techniques for the timely detection of outdated dependencies to prevent security issues. We aggregated outdated dependencies from the npm ecosystem and the Open Source Vulnerabilities (OSV) dataset to construct the Outdated Dependencies Vulnerability Supply Chain. Our analysis suggested that outdated dependencies in a central position within the dependency chain significantly affect downstream components. Using clustering algorithms, we discovered that outdated packages are concentrated in the development tools area, which has a significant security risk from injection flaws (CWE-707) vulnerabilities. Utilizing model-based feature importance techniques, we predict the outdatedness of dependencies in the open source supply chain, with key indicators including the average duration of open issues and pull requests. The findings offer software developers and maintainers a foundation to identify current issues and security vulnerabilities within outdated dependencies, facilitating mitigation efforts through the open source software supply chain.

The building industry plays a significant role in global carbon emissions, contributing nearly half of the world’s greenhouse gas emissions during both construction and operation. Within the framework of the “double-low” strategy, addressing energy conservation, emission reduction, and climate adaptation in buildings has become a crucial area of research and practice. In northern China, vernacular dwellings have historically developed passive strategies for climate adaptation; however, their quantified thermal performance has not been thoroughly studied. This research focuses on single-story courtyard vernacular dwellings built in the 1990s, which are inspired by historical Siheyuan forms in Shatun Village, located in Handan, Hebei Province. The study specifically examines their thermal performance during the summer and the relationship between this performance and climate design strategies. To understand how building layout, envelopes, materials, and courtyard landscape design influence the microclimate, six measurement points were established within each dwelling to continuously collect environmental data, including air temperature, humidity, and wind speed. The RayMan model was used to calculate the mean radiant temperature (Tmrt) and physiological equivalent temperature (PET), with subsequent statistical analysis conducted using Origin Pro. The results showed that sustainable design strategies—such as high building envelopes, shaded vegetation, and low-albedo materials—contributed to maintaining a stable microclimate, with over 70% of daytime PET values remaining within a comfortable range. Night-time cooling and the increased humidity from courtyard vegetation significantly enhance thermal resilience. It is important to distinguish this from ambient humidity, which can hinder human evaporative cooling and increase heat stress during extreme heat. This research demonstrates that vernacular dwellings can achieve thermal comfort without relying on mechanical cooling systems. These findings provide strong empirical support for incorporating passive, courtyard-based climate strategies in contemporary rural housing worldwide, contributing to low-carbon and climate-resilient development beyond regional contexts.

Cisplatin-based chemotherapy remains a mainstay for the treatment of bladder cancer (BLCA); however, its clinical efficacy is frequently compromised by the emergence of chemoresistance, which leads to poor patient outcomes. Although known mechanisms—such as alterations in drug efflux, DNA repair, and key signaling pathways—have been implicated, they fail to fully explain the clinical complexity of cisplatin resistance, indicating that additional molecular drivers remain undiscovered. Keratin 14 (KRT14), an intermediate filament protein associated with aggressive BLCA subtypes, is consistently upregulated in cisplatin-resistant tumors, yet its precise mechanistic role in resistance remains unclear. In this study, we elucidate the functional contribution of KRT14 to cisplatin resistance in BLCA using patient-derived tissues, established cell lines, xenograft mouse models, and a suite of molecular interaction assays. Our results demonstrate that KRT14 is significantly upregulated in BLCA tissues, correlates with poor clinical prognosis, and functionally drives cisplatin resistance both in vitro and in vivo. Mechanistically, we identify a novel and direct interaction between KRT14 and the translation initiation factor eIF4H, specifically through the N-terminal Head domain of KRT14. This interaction modulates the association of eIF4H with the core eIF4F complex, thereby selectively promoting the translation of Acyl-CoA Oxidase 2 (ACOX2) mRNA through its 5’ untranslated region. We further show that ACOX2 is essential for mediating the effects of KRT14 on lipid metabolism, cell proliferation, survival, and ultimately, cisplatin resistance. Collectively, our findings reveal that KRT14 contributes to chemoresistance in BLCA not only via its structural roles but also by directly regulating translational machinery through eIF4H, leading to upregulation of the metabolic enzyme ACOX2. The newly defined KRT14–eIF4H–ACOX2 axis orchestrates lipid metabolic reprogramming and cell survival, underscoring the KRT14–eIF4H interface as a promising therapeutic target for overcoming cisplatin resistance in BLCA.

Abstract Background To determine the effects of negative heel shoes on perceived pain and knee biomechanical characteristics of runners with patellofemoral pain (PFP) during running. Methods Sixteen runners with PFP ran in negative (−11 mm drops) and positive (5 mm drops) heel shoes while visual analog scale (VAS) scores, retroreflective markers, and ground reaction force were acquired by applying a 10‐cm VAS, infrared motion capture system, and a three‐dimensional force plate. Knee moment, patellofemoral joint stress (PFJS), and other biomechanical parameters during the stance phase were calculated based on inverse dynamics and a biomechanical model of the patellofemoral joint. Results The foot inclination angle, peak PFJS during the stance phase, patellofemoral joint reaction force, knee extension moment, and quadriceps force at the time of peak PFJS of runners with PFP in negative heel shoes were lower than that in positive heel shoes, no significant difference was found in VAS scores, knee flexion angle, patellofemoral contact area, and quadriceps moment arm at the time of peak PFJS. Conclusions Compared to positive heel shoes, running in negative heel shoes decreases peak PFJS in runners with PFP, which may decrease patellofemoral joint loading, thus reducing the possibility of further development of PFP. Trail Registration Sports Science Experiment Ethics Committee of Beijing Sport University. 2023095H, April 18, 2023 (prospectively registered).

Zeolitic imidazolate frameworks (ZIFs) are a series of materials composited by metal ions and organic ligands with high specific surface area, which might be great precursors to produce metal oxides by calcination for gas sensor application. However, Zn-ZIF (ZIF-8) is hard to transform as ZnO in air and keeping the unique framework simultaneously. In this work, Fe2+ was introduced into the metal node to replace a part of Zn2+ ions, and it could be oxidized as Fe3+ in the calcination to facilitate the oxidation process of the 2-methylimdazole ligands to give Fe-ZnO complex shell with high specific surface area (108 m2/g) and abundant oxygen vacancies (48%). The micro electro mechanical systems (MEMS) sensor based on the 6%-Fe-ZnO complex shell performed outstanding gas sensing properties to the low-concentration acetone vapor, including high response (ΔR/Rg = 11.2 to 5 ppm acetone), superior selectivity (Sacetone/Sethanol = 5.6) and fast response speed (τres = 2.6 s). This work not only provided the research of an exceptional acetone MEMS sensor, but also induced a strategy to produce metal oxide derived from ZIFs with complex structures for the universal synthesis methodology.

Precisely tailoring the surface electronic state of catalysts to realize the optimal design of adsorption sites is essential to the surface-related gas-sensing reaction. Herein, based on both molecular orbital theory and p-band models, we develop a brilliant surface oxygen-injected method to simultaneously enhance the overlap of energylevel alignment (ELA) and reduce the anti-bonding filling (ABF) level between surface Bi p-band and adsorbed NO 2 molecule, leading to an optimal NO 2 adsorption mode and sensing performance. By controlling the oxygen permeation concentrations, the weak-oxidized Bi 2 S 3-200 catalysts with ordered core/disordered shell configuration exhibit excellent NO 2 gas sensitivity (12.5 % to 1 ppm) and low experimental detection limit (100 ppb), surpassing that of most reported NO 2 sensors. Ex situ XPS characterizations further demonstrate that the weakoxidized amorphous Bi species can serve as active adsorption centers to alter the electron transfer path in NO 2 atmosphere. Finally, through inserting flexible MEMS sensors array into multifunctional wireless sensing device, the Bi 2 S 3-200 sensors can realize real-time NO 2 /temperature/humidity monitoring and cloud data transmission at room temperature, which thereby pave the way for the development of crop health monitor and precision agriculture.

Viscoelasticity is a crucial property of cells, which plays an important role in label-free cell characterization. This paper reports a model-fitting-free viscoelasticity calculation method, correcting the effects of frequency, surface adhesion and liquid resistance on AFM force-distance (FD) curves. As demonstrated by quantifying the viscosity and elastic modulus of PC-3 cells, this method shows high self-consistency and little dependence on experimental parameters such as loading frequency, and loading mode (Force-volume vs. PeakForce Tapping). The rapid calculating speed of less than 1ms per curve without the need for a model fitting process is another advantage. Furthermore, this method was utilized to characterize the viscoelastic properties of primary clinical prostate cells from 38 patients. The results demonstrate that the reported characterization method a comparable performance with the Gleason Score system in grading prostate cancer cells, This method achieves a high average accuracy of 97.6% in distinguishing low-risk prostate tumors (BPH and GS6) from higher-risk (GS7-GS10) prostate tumors and a high average accuracy of 93.3% in distinguishing BPH from prostate cancer.

Hydrogen sulfide is a kind of widely distributed toxic gas, therefore, it is strongly necessary to conduct highly selective detection at lower concentrations. However, due to the high affinity of sulfur elements to transition metals, hydrogen sulfide sensors based on metal oxide semiconductors often have slower recovery and poor selectivity. In this work, Fe-NiOx nanotubes with unique structure and high specific surface area (175 m2/g) were synthesized by using anodized aluminum oxide templates for detecting sub-ppm H2S. The gas sensor based on 2 wt% Fe-doped NiOx nanotube performed exceptional gas sensing properties to 800 ppb H2S, featured with the high sensitivity (|Delta R/Ra| = 5.24), excellent selectivity (SH2S/SNO2 = 8.3) and fast response/recovery speed (3.2/ 8.1 s) at 270 degrees C. Density function theory calculation proved the critical function of Fe3+ to lower the adsorption enthalpy of H2S and changing the selectivity. This work not only provided a superior ppb-level H2S gas sensor, but also induced a synthesis strategy to enhance the gas sensing properties.

In the winter environment of cold regions, the visual perception of inter-house leisure space in settlements has an important impact on the physical and mental health of residents. The study used an eye-tracking device combined with a semantic differential (SD) questionnaire to screen the visual attention elements and attributes of winter residents in cold regions, including the spatial aspect ratio (D/H), residential elevation saturation (RES), and the percentage of lawn in the field of view (POL). Orthogonal experiments were established in an immersive virtual environment to reveal the influence mechanism of visual perceptual environmental factors of cold regions settlements on residents' leisure space evaluation in winter environment. The study trained and compared four visual perception machine learning agent models, combining genetic algorithms (GA) with k-nearest neighbor algorithms (KNN), resulting in optimized threshold ranges of D/H: 2.22-2.54, RES: 68.47-82.34, and POL: 10%-14% in winter environments.