[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-655891-105":59,"doc-detail-655891-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","text-to-automata-diagrams-comparing-tikz-code-generation-with-direct-image-synthesis","Text to Automata Diagrams - Comparing TikZ Code Generation with Direct Image Synthesis","","Diagrams are central to computer science education, especially in automata and formal languages, where hand-drawn student sketches vary in structure, layout, and correctness. This study evaluates whether vision-language and large language models can transform scanned diagrams into accurate textual descriptions and then into TikZ representations. Image-derived descriptions are reviewed and revised by humans, after which both raw and revised outputs are used to generate TikZ code for compilation and visual comparison. Results show direct image descriptions are often incorrect, and human correction improves description and final diagram accuracy.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":30,"@type":76,"position":81},"https://docshare.wps.com/document/technology/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/text-to-automata-diagrams-comparing-tikz-code-generation-with-direct-image-synthesis/655891/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/text-to-automata-diagrams-comparing-tikz-code-generation-with-direct-image-synthesis/655891.png","ImageObject",300,407,{"name":92,"@type":93},"Indoniesian Boy","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-02","2026-10-25",true,{"@type":102,"interactionType":103,"userInteractionCount":29},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What reconstruction pipeline does the paper evaluate for student automata sketches?","Question",{"text":112,"@type":113},"It uses scanned diagrams to generate textual descriptions with a vision-language model, optionally revises them with human reviewers, then feeds the descriptions into a large language model to produce TikZ code. The compiled TikZ diagrams are visually compared to the original sketches.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"Why do the authors include human review in the workflow?",{"text":117,"@type":113},"Because descriptions generated directly from images using vision-language models are often incorrect. Human correction revises errors so downstream LLM-generated TikZ code better matches the student’s original diagram.",{"name":119,"@type":110,"acceptedAnswer":120},"What comparison results are reported across the three study stages?",{"text":121,"@type":113},"The study compares semantic differences between raw and edited descriptions, the diagram quality produced from raw versus edited descriptions, and the diagram quality produced from TikZ code generated from raw versus edited descriptions. Overall, diagrams derived from edited descriptions are more accurate.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},655891,1793073174,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":29,"category_name":30,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":29,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":41},962090760608,"https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c","Paper ID \\#54216  \nText to Automata Diagrams: Comparing TikZ Code Generation with Direct Image Synthesis  \nEthan Young, Marshall University  \nAniruddha Maiti  \nDr. Maiti is an Assistant Professor of Mathematics, Engineering and Computer Science at West Virginia State University. His research interest is in AI, especially in traditional machine learning, multimodal learning, large language and vision models.  \nDr. Ananya Jana, Marshall University  \nZichun Wang  \nSatya Sri Rajiteswari Nimmagadda, Marshall University  \n©American Society for Engineering Education, 2026  \nText to Automata Diagrams: Comparing TikZ Code Generation with Direct Image Synthesis  \nEthan Young 1 , Zichun Wang2 , Aiden Taylor 1 , Chance Jewell 1 , Julian Myers 1 , Satya Sri Rajiteswari Nimmagadda 1 , Anthony White II2 , Aniruddha Maiti2 ,  \nAnanya Jana 1  \n1Marshall University 2West Virginia State University  \nAbstract  \nDiagrams are widely used in teaching computer science courses. They are useful in subjects such as automata and formal languages, data structures, etc. These diagrams, often drawn by students during exams or assignments, vary in structure, layout, and correctness. This study examines whether current vision-language and large language models can process such diagrams and produce accurate textual and digital representations. In this study, scanned studentdrawn diagrams are used as input. Then, textual descriptions are generated from these images using a vision-language model. The descriptions are checked and revised by human reviewers to make them accurate. Both the generated and the revised descriptions are then fed to a large language model to generate TikZ code. The resulting diagrams are compiled and then evaluated against the original scanned diagrams. We found descriptions generated directly from images using vision-language models are often incorrect and human correction can substantially improve the quality of vision language model generated descriptions. This research can help computer science education by paving the way for automated grading and feedback and creating more accessible instructional materials.  \nIntroduction  \nAutomata diagrams appear in undergraduate computer science courses that cover formal languages and computation. Students use hand-drawn sketches to take notes, to solve problems, to explain processes, or to show how systems behave. These diagrams describe the structure of finite automata through labeled states and transitions. Students draw these sketches by hand during exams, quizzes, and assignments. The drawings are often informal and vary in layout, notation, and correctness. Vision language models can process scanned diagrams and generate text. The large language models (LLM) can take structured text and produce LaTeX TikZ code. TikZ is a typesetting language commonly used for diagrams in instructional material. The sequence from diagram to description to TikZ can be used to reconstruct a student’s diagram in a clean format. This reconstruction depends on both the accuracy of the generated text and the ability of the model to follow the diagram conventions used in computer science theory subjects such as automata.  \nThis paper studies the reliability of such a reconstruction process. The experiments use a collection of student-drawn diagrams. A description for each image is generated by a visionlanguage model (GPT-4o) . We found that such descriptions are not always accurate. For this reason, human review is performed to revise the descriptions to correct errors in these descriptions. Both types of descriptions are used as input to a large language model to generate corresponding TikZ code. The TikZ output is compiled and visually compared with the original sketch.  \nThe study includes three comparisons. First, we compared the semantic content of the raw and edited descriptions using similarity scores to understand how much they differ. Second, we  \ncompared the diagrams generated from raw descriptions and ","cbCaijxqVawuvzEh","https://ap.wps.com/l/cbCaijxqVawuvzEh","pdf",1106068,12,"English","# Abstract\n# Introduction\n# Related Work\n# Dataset Preparation\n## Image scan collection","[{\"question\":\"What reconstruction pipeline does the paper evaluate for student automata sketches?\",\"answer\":\"It uses scanned diagrams to generate textual descriptions with a vision-language model, optionally revises them with human reviewers, then feeds the descriptions into a large language model to produce TikZ code. The compiled TikZ diagrams are visually compared to the original sketches.\"},{\"question\":\"Why do the authors include human review in the workflow?\",\"answer\":\"Because descriptions generated directly from images using vision-language models are often incorrect. Human correction revises errors so downstream LLM-generated TikZ code better matches the student’s original diagram.\"},{\"question\":\"What comparison results are reported across the three study stages?\",\"answer\":\"The study compares semantic differences between raw and edited descriptions, the diagram quality produced from raw versus edited descriptions, and the diagram quality produced from TikZ code generated from raw versus edited descriptions. Overall, diagrams derived from edited descriptions are more accurate.\"}]","Text to Automata Diagrams - Comparing TikZ Code Generation with Direct Image Synthesis | PDF",1792964619]