ControlNet插件在AI生成图像中的控制应用分析

运用AI绘图工具Stable Diffusion生成图像的过程中,虽然可以用提示词来描述画面,但生成图像的不确定性很大。为了生成满意的图像,可以使用Stable Diffusion的插件ControlNet对图像生成进行控制。本文分析了用ControlNet对图像生成过程的控制方法,并运用该方法在具体案例中实现了在生成图像过程中对画面构图的控制。

基于AI图像生成技术的家具设计研究

简述了AI图像生成技术及其在家具设计中的应用优势,提出AI图像生成技术在家具设计中的一般流程。基于Stable Diffusion AI设计平台,依据“文本指令”进行家具设计尝试,以包豪斯风格家具设计实践来验证AI图像生成技术辅助生成家具设计效果图的高效性。研究表明:AI图像生成技术能合理运用于家具设计创意阶段,对缩短家具设计周期、提升家具设计质量具有实际意义。

基于开源大模型的短视频与直播AI实训系统的设计与实现

提出了一种基于开源大模型的短视频与直播AI实训系统设计方案,旨在通过集成Llama-3、Stable Diff usion、Stable Video Diff usion等先进技术,构建一个集教学与实践于一体的智能化平台,促进新媒体人才的培养。系统设计覆盖账号定位分析、剧本策划、分镜头生成至视频合成的全链路,利用ComfyUI提供用户友好的操作界面,NestJS实现高效的服务器管理,以及MySQL和MongoDB支持数据存储。经过功能测试与用户体验评估,所设计的系统有效提升了学生的实践技能与内容创作水平,具有良好的应用前景。未来发展方向包括模型的持续优化、技术升级,以及加强与产业界的合作,以更好地适配行业需求,开拓新媒体教育的未来空间。

人工智能在电影制作领域的应用探究——以分镜头脚本生成工具为例

随着人工智能技术的发展与应用,以Midjourney、Stable Diffusion为代表的人工智能(Artificial Intelligence,AI)图像生产工具凭借其高效率、低成本、易学习等特点,在图像绘画领域体现出较强竞争力。分镜头脚本的绘制与设计是电影前期制作的重要环节,本文着重研究将人工智能生成内容(Artificial Intelligence Generated Content,AIGC)技术应用到电影分镜头脚本的绘制当中,通过比较AIGC生成的分镜头脚本与人类创作的差异,探讨AIGC在提高创作效率、内容创新性等方面的潜在优势,此外,还将探讨AIGC技术在数字电影制作上所面临的挑战,如创造性和情感表达的限制,以及与人类创作者的合作模式,以探索实现人工智能领域分镜头脚本生成的新的可能性。

从特征辨识到图像生成:基于AIGC范式的苗族服饰设计

苗族服饰是中国民族服饰体系中重要的组成部分,在商品化发展中开辟了广阔的市场,但产品同质化现象将阻碍民族文化的进一步发展,人工智能生成内容(AIGC)的发展为传统服饰设计提供了新的可能。文章基于苗族服饰现代化的需要,通过搭建多模型组合流程,为苗族服饰成衣化设计提供灵感样本。实验结果表明,以服饰辨识性为基础,采用Lora训练方式,仅需少量样本即可实现苗族服饰特征在Stable Diffusion生成内容中的迁移。将深度学习引入非遗服饰现代化的进程中,旨在助力非遗服饰朝着成衣化、多样化、大众化方向发展,以数字化方式推动民族服饰转型升级。

AIGC视野下的浙北民居风格研究路径初探

文章以Stable Diffusion为工具,以通过机器学习特定建筑风格图像并输出结果的方式,尝试对浙北民居的建筑风格在AIGC视野下的研究路径进行探索。文章建立了浙北民居与现代住宅图像的数据集并进行训练,得到该语境下的浙北民居风格模型,接着通过批量单独生成与混合生成,对生成的结果进行分类统计。最终发现了该模型在建筑风格模拟方面具有不同于人类直觉判断的抽象提取能力,进而概括了该类工具下生成结果的可能性与局限性。

基于美学梯度法的人工智能风格化绘画系统

当前Stable diffusion等人工智能绘画模型在绘画时难以直接控制图像风格,同时风格模型训练仅针对单种风格。针对该问题,提出了一种基于美学梯度法的人工智能风格化绘画系统,以实现多种图像风格的控制和融合,并提供更加便捷的图像创作体验。收集并分析网络用户数据,结合问卷得到用户对图像风格的感性需求;根据感性需求收集各风格图像数据得到对应的风格图像训练集。使用梯度下降算法计算风格化文本编码器的权重,实现生成图像风格化的效果。通过可用性测试对比用户对该系统与传统人工智能绘画系统产出图像的风格满意程度,结果表明:人工智能风格化绘画系统的平均满意度相较传统人工智能绘画系统提升23%,表明人工智能风格化绘画系统在图像风格生成上具有更好的效果,可满足用户对图像风格的需求。该人工智能风格化绘画系统可以更便捷地实现图像风格调整,允许用户直观选择不同风格的权重,便捷使用一种或多种风格,能够有效满足用户对图像风格设计的需求。

Geochemistry and origins of hydrogen-containing natural gases in deep Songliao Basin,China:Insights from continental scientific drilling

The different reservoirs in deep Songliao Basin have non-homogeneous lithologies and include multiple layers with a high content of hydrogen gas.The gas composition and stable isotope characteristics vary significantly,but the origin analysis of different gas types has previously been weak.Based on the geochemical parameters of gas samples from different depths and the analysis of geological settings,this research covers the diverse origins of natural gas in different strata.The gas components are mainly methane with a small amount of C_(2+),and non-hydrocarbon gases,including nitrogen(N_(2)),hydrogen(H_(2)),carbon dioxide(CO_(2)),and helium(He).At greater depth,the carbon isotope of methane becomes heavier,and the hydrogen isotope points to a lacustrine sedimentary environment.With increasing depth,the origins of N_(2)and CO_(2)change gradually from a mixture of organic and inorganic to inorganic.The origins of hydrogen gas are complex and include organic sources,water radiolysis,water-rock(Fe^(2+)-containing minerals)reactions,and mantle-derived.The shales of Denglouku and Shahezi Formations,as source rocks,provide the premise for generation and occurrence of organic gas.Furthermore,the deep faults and fluid activities in Basement Formation control the generation and migration of mantle-derived gas.The discovery of a high content of H_(2)in study area not only reveals the organic and inorganic association of natural-gas generation,but also provides a scientific basis for the exploration of deep hydrogen-rich gas.

基于Stable Diffusion的虚拟人形象预设计的应用与研究

对当前AIGC在虚拟形象预设计方面的现状及影响进行了分析和探讨。以Stable Diffusion为例,详细介绍了工程构建和实现,对相关模块的作用、运行环境、使用方法及其指令等多个方面进行了综合叙述、分析和探讨,针对使用不同采样方法、不同采样参数及不同训练模型生成图片效果的优劣进行了说明及展示。随后,通过项目实例,完整地展示了人物形象预设计的过程。最后,对AIGC等新技术可能带来的社会影响进行了预测和总结。

Rising utilization of stable isotopes in tree rings for climate change and forest ecology

Analyses of stable isotopes(C,O,H)in tree rings are increasingly important cross-disciplinary programs.The rapid development in this field documented in an increasing number of publications requires a comprehensive review.This study includes a bibliometric analysis-based review to better understand research trends in tree ring stable isotope research.Overall,1475 publications were selected from the Web of Science Core Collection for 1974-2023.The findings are that:(1)numbers of annual publications and citations increased since 1974.From 1974 to 1980,there were around two relevant publications per year.However,from 2020 to 2022,this rose sharply to 109 publications per year.Likewise,average article citations were less than four per year before 1990,but were around four per article per year after 2000;(2)the major subjects using tree ring stable isotopes include forestry,geosciences,and environmental sciences,contributing to 42.5%of the total during 1974-2023;(3)the top three most productive institutions are the Chinese Academy of Sciences(423),the Swiss Federal Institute for Forest,Snow and Landscape Research(227),and the University of Arizona(204).These achievements result from strong collaborations;(4)review papers,for example,(Dawson et al.,Annu Rev Ecol Syst 33:507-559,2002)and(McCarroll and Loader,Quat Sci Rev 23:771-801,2004),are among the most cited,with more than 1000 citations;(5)tree ring stable isotope studies mainly focus on climatology and ecology,with atmospheric CO_(2) one of the most popular topics.Since 2010,precipitation and drought have received increasing attention.Based on this analysis,the research stages,key findings,debated issues,limitations and direc-tions for future research are summarized.This study serves as an important attempt to understand the progress on the use of stable isotopes in tree rings,providing scientific guid-ance for young researchers in this field.

Method to measure tree-ring width,density,elemental composition,and stable carbon and oxygen isotopes using one sample

Tree-ring width(RW),density,elemental com-position,and stable carbon and oxygen isotope(δ^(13)C,δ^(18)O)are widely used as proxies to assess climate change,ecology,and environmental pollution;however,a specific pretreat-ment has been needed for each proxy.Here,we developed a method by which each proxy can be measured in the same sample.First,the sample is polished for ring width meas-urement.After obtaining the ring width data,the sample is cut to form a 1-mm-thick wood plate.The sample is then mounted in a vertical sample holder,and gradually scanned by an X-ray beam.Simultaneously,the count rates of the fluorescent photons of elements(for chemical characteriza-tion)and a radiographic grayscale image(for wood density)are obtained,i.e.the density and the element content are obtained.Then,cellulose is isolated from the 1-mm wood plate by removal of lignin,and hemicellulose.After producing this cellulose plate,cellulose subsamples are separated by knife under the microscope for inter-annual and intra-annual stable carbon and oxygen isotope(δ^(13)C,δ^(18)O)analysis.Based on this method,RW,density,elemental composition,δ^(13)C,and δ^(18)O can be measured from the same sample,which reduces sample amount and treatment time,and is helpful for multi-proxy comparison and combination research.

Software Defect Prediction Method Based on Stable Learning

The purpose of software defect prediction is to identify defect-prone code modules to assist software quality assurance teams with the appropriate allocation of resources and labor.In previous software defect prediction studies,transfer learning was effective in solving the problem of inconsistent project data distribution.However,target projects often lack sufficient data,which affects the performance of the transfer learning model.In addition,the presence of uncorrelated features between projects can decrease the prediction accuracy of the transfer learning model.To address these problems,this article propose a software defect prediction method based on stable learning(SDP-SL)that combines code visualization techniques and residual networks.This method first transforms code files into code images using code visualization techniques and then constructs a defect prediction model based on these code images.During the model training process,target project data are not required as prior knowledge.Following the principles of stable learning,this paper dynamically adjusted the weights of source project samples to eliminate dependencies between features,thereby capturing the“invariance mechanism”within the data.This approach explores the genuine relationship between code defect features and labels,thereby enhancing defect prediction performance.To evaluate the performance of SDP-SL,this article conducted comparative experiments on 10 open-source projects in the PROMISE dataset.The experimental results demonstrated that in terms of the F-measure,the proposed SDP-SL method outperformed other within-project defect prediction methods by 2.11%-44.03%.In cross-project defect prediction,the SDP-SL method provided an improvement of 5.89%-25.46% in prediction performance compared to other cross-project defect prediction methods.Therefore,SDP-SL can effectively enhance within-and cross-project defect predictions.

Observer-based robust high-order fully actuated attitude autopilot design for spinning glide-guided projectiles

This paper investigates the design of an attitude autopilot for a dual-channel controlled spinning glideguided projectile(SGGP),addressing model uncertainties and external disturbances.Based on fixed-time stable theory,a disturbance observer with integral sliding mode and adaptive techniques is proposed to mitigate total disturbance effects,irrespective of initial conditions.By introducing an error integral signal,the dynamics of the SGGP are transformed into two separate second-order fully actuated systems.Subsequently,employing the high-order fully actuated approach and a parametric approach,the nonlinear dynamics of the SGGP are recast into a constant linear closed-loop system,ensuring that the projectile's attitude asymptotically tracks the given goal with the desired eigenstructure.Under the proposed composite control framework,the ultimately uniformly bounded stability of the closed-loop system is rigorously demonstrated via the Lyapunov method.Validation of the effectiveness of the proposed attitude autopilot design is provided through extensive numerical simulations.

Differential response of radial growth and δ^(13)C in Qinghai spruce(Picea crassifolia) to climate change on the southern and northern slopes of the Qilian Mountains in Northwest China

Tree radial growth can have significantly differ-ent responses to climate change depending on the environ-ment.To elucidate the effects of climate on radial growth and stable carbon isotope(δ^(13)C)fractionation of Qing-hai spruce(Picea crassifolia),a widely distributed native conifer in northwestern China in different environments,we developed chronologies for tree-ring widths and δ^(13)C in trees on the southern and northern slopes of the Qilian Mountains,and analysed the relationship between these tree-ring variables and major climatic factors.Tree-ring widths were strongly influenced by climatic factors early in the growing season,and the radial growth in trees on the northern slopes was more sensitive to climate than in trees on the southern.Tree-ring δ^(13)C was more sensitive to climate than radial growth.δ^(13)C fractionation was mainly influenced by summer temperature and precipitation early in the growing season.Stomatal conductance more strongly limited stable carbon isotope fractionation in tree rings than photosynthetic rate did.The response between tree rings and climate in mountains gradually weakened as climate warmed.Changes in radial growth and stable carbon isotope fractionation of P.crassifolia in response to climate in the Qilian Mountains may be further complicated by continued climate change.

基于生成式AI的人工智能在建筑设计中的应用探究

近两年间人工智能(AI)相关技术逐步由顶端智能产业下放,走进越来越多诸如建筑等传统行业中。把握住与新型技术相结合的机会是建筑设计领域一种可探索的道路。文章将以生成式AI为例,从应用领域、应用模式、方案生成等视角,探究人工智能在建筑方案设计领域可以为建筑师提供的参考和帮助,并提出其中的缺陷与局限性,为今后人工智能在建筑设计领域更加深入、系统地应用提供参考。

基于提示词优化的AIGC汉长安城遗址儿童绘本应用研究

以汉长安城遗址中椒房殿为测试对象,探究和优化Stable Diffusion中的提示词输入策略生成椒房殿复原设想图,并且依据该输入逻辑制作出儿童绘本素材图。研究将AI绘画、汉文化和儿童绘本创新性的融合,对于AI艺术应用和儿童美育教育实践具有可供参考的推广价值。

从Make-A-Video到Sora:AI视频生成技术的进步与挑战

随着人工智能技术的飞速发展,AI视频生成技术已成为研究和应用的热点。从Meta的MakeA-Video, Runway AI的Runway Gen-2,Stability AI的Stable Video Diffusion,到Google的Lumiere,再到OpenAI的Sora,每一个模型的推出都不仅代表了AI视频生成技术的进步,也带来了新的挑战。回顾了这些关键AI视频模型的原理和特点,并对比它们之间的优势和不足,探讨了AI视频生成技术面临的主要挑战,展望了未来的发展方向。

Decline in tree-ring growth of Picea mongolica and its intra-annual eco-physiological responses to drought and CO_(2) enrichment in semi-arid China

Episodes of drought-induced decline in tree growth and mortality are becoming more frequent as a result of climate warming and enhanced water stress in semi-arid areas.However,the ecophysiological mechanisms under-lying the impact of drought on tree growth remains unre-solved.In this study,earlywood and latewood tree-ring growth,δ^(13)C,andδ^(18)O chronologies of Picea mongolica from 1900 to 2013 were developed to clarify the intra-and inter-annual tree-ring growth responses to increasingly fre-quent droughts.The results indicate that annual basal area increment residuals(BAI_(res)),which removed tree age and size effects,have significantly decreased since 1960.How-ever,the decreasing trend of earlywood BAI_(res) was higher than that of latewood.Climate response analysis suggests that the dominant parameters for earlywood and latewood proxies(BAI_(res),δ^(13)C andδ^(18)O)were drought-related climate variables(Palmer drought severity index,temperature,rela-tive humidity,and vapor pressure deficit).The most signifi-cant period of earlywood and latewood proxies’responses to climate variables were focused on June-July and July-August,respectively.BAI_(res),andδ^(13)C were significantly affected by temperature and moisture conditions,whereasδ^(18)O was slightly affected.Decreasing stomatal conduct-ance due to drought outweighed the influence of increasing CO_(2) on intrinsic water use efficiency(iWUE),and ultimately led to a decline in BAI_(res).Compared to latewood,the faster decreasing BAI_(res) and smaller increasing iWUE of early-wood suggested trees were more vulnerable to water stress in the early growing season.Our study provides insights into the inter-and intra-annual mechanisms of tree-ring growth in semi-arid regions under rising CO_(2) and climate change.

Determinant variables on the isotopic values of particulate organic matter in a neotropical floodplain

Particulate organic matter(POM)is an important energy source for aquatic consumers,understanding its origin and composition is essential for understanding the energetic dynamics of aquatic environments.The aim of this study was to analyze the relationship between POM and phytoplankton(isotopic values and chlorophyll concentration)and abiotic variables during dry and rainy seasons.Sampling was conducted in rivers and lagoons in the floodplain of the Upper ParanáRiver.We found a greater difference in ^(δ13)C values of POM between sampling points than between seasons,indicating that the composition of regional sources influences the composition of POM more than dry and rainy seasons.In addition,the concentration of chlorophyll during the dry season was positively correlated with ^(δ13)C values during that rainy period.Additionally,we found a relationship between factors limiting the growth of phytoplankton and ^(δ13)C values of POM,such as phosphate ions,indicating that variables that regulate phytoplankton growth tend to influence the composition of POM in river floodplains.Therefore,maintaining the variables that regulate the phytoplankton community is of fundamental importance for the composition of POM,an important energy source in aquatic environments.

Exploring the Latest Applications of OpenAI and ChatGPT: An In-Depth Survey

OpenAI and ChatGPT, as state-of-the-art languagemodels driven by cutting-edge artificial intelligence technology,have gained widespread adoption across diverse industries. In the realm of computer vision, these models havebeen employed for intricate tasks including object recognition, image generation, and image processing, leveragingtheir advanced capabilities to fuel transformative breakthroughs. Within the gaming industry, they have foundutility in crafting virtual characters and generating plots and dialogues, thereby enabling immersive and interactiveplayer experiences. Furthermore, these models have been harnessed in the realm of medical diagnosis, providinginvaluable insights and support to healthcare professionals in the realmof disease detection. The principal objectiveof this paper is to offer a comprehensive overview of OpenAI, OpenAI Gym, ChatGPT, DALL E, stable diffusion,the pre-trained clip model, and other pertinent models in various domains, encompassing CLIP Text-to-Image,education, medical imaging, computer vision, social influence, natural language processing, software development,coding assistance, and Chatbot, among others. Particular emphasis will be placed on comparative analysis andexamination of popular text-to-image and text-to-video models under diverse stimuli, shedding light on thecurrent research landscape, emerging trends, and existing challenges within the domains of OpenAI and ChatGPT.Through a rigorous literature review, this paper aims to deliver a professional and insightful overview of theadvancements, potentials, and limitations of these pioneering language models.