[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-292255-105":59,"doc-detail-292255-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","a-novel-dataset-for-video-based-neurodivergent-classification-leveraging-extra-stimulatory-behavior","A Novel Dataset for Video-Based Neurodivergent Classification Leveraging Extra-Stimulatory Behavior","","Facial expressions and actions vary in intensity across individuals, especially among neurodivergent people, influencing health, communication, and sensory processing. This work presents the Video ASD dataset with convolutional and attention-map frame features designed to advance ASD classification. Unlike MRI-based approaches requiring specialized expensive equipment, the method uses an affordable GPU, standard computing, and a video camera for inference. Results show effective generalization across children’s distinct movements, and foundation-model tests assess how movement noise impacts performance and data needs.",{"@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":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/a-novel-dataset-for-video-based-neurodivergent-classification-leveraging-extra-stimulatory-behavior/292255/",{"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/a-novel-dataset-for-video-based-neurodivergent-classification-leveraging-extra-stimulatory-behavior/292255.png","ImageObject",300,407,{"name":92,"@type":93},"Stanley","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-10-30","2026-09-17",true,{"@type":102,"interactionType":103,"userInteractionCount":34},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What problem does the Video ASD dataset target?","Question",{"text":112,"@type":113},"It targets ASD-related behavior classification from video, capturing facial expressions and head-pose-linked actions under external stimuli, with special attention to neurodivergent variability.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How are video inputs represented in the dataset and model?",{"text":117,"@type":113},"Each dataset instance contains a sequence of frame features, alongside labels (yes/no) and timed text descriptions; the model processes 32 frames per video for spatial feature extraction and temporal decoding.",{"name":119,"@type":110,"acceptedAnswer":120},"What experimental setup and performance results are reported?",{"text":121,"@type":113},"The framework is implemented in PyTorch and trained with 5-fold cross-validation using AdamW and cosine annealing, with face alignment and frame filtering; the reported test accuracy is 81.48% with an F1 score of 0.7289.","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},292255,1789649409,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":34,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":14,"language":139,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":140,"faqs":141,"seo_title":142,"seo_description":67,"update_tm":129,"read_time":24,"preview_page_count":4},2336477405376,"https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0","A Novel Dataset for Video-Based Neurodivergent Classification Leveraging Extra-Stimulatory  \nBehavior  \nManuel Serna-Aguilera  \nDept. of EECS University of Arkansas Fayetteville, United States [mserna@uark.edu](mserna@uark.edu)  \nXuan Bac Nguyen  \nDept. of EECS University of Arkansas Fayetteville, United States [xnguyen@uark.edu](xnguyen@uark.edu)  \nHan-Seok Seo  \nDept. of Food Science University of Arkansas Fayetteville, United States [hanseok@uark.edu](hanseok@uark.edu)  \nKhoa Luu  \nDept. of EECS University of Arkansas Fayetteville, United States [khoaluu@uark.edu](khoaluu@uark.edu)  \narXiv :2409 .04598v2 [ cs .CV] 22 Aug 2025  \nAbstract—Facial expressions and actions differ among different individuals at varying degrees of intensity given responses to external stimuli, particularly among those that are neurodivergent. Such behaviors affect people in terms of overall health, communication, and sensory processing. Deep learning can be responsibly leveraged to improve productivity in addressing this task, and help medical professionals to accurately understand such behaviors. In this work, we introduce the Video ASD dataset—a dataset that contains video frame convolutional and attention map feature data—to foster further progress in the task of ASD classification. Unlike many recent studies in ASD classification with MRI data, which require expensive specialized equipment, our method utilizes a powerful but relatively affordable GPU, a standard computer setup, and a video camera for inference. Results show that our model effectively generalizesand understands key differences in the distinct movements of the children. Additionally, we test foundation models on this data to showcase how movement noise affects performance and the need for more data and more complex labels.  \nIntroduction and Related Work. Deep learning has seen success in various deep learning tasks [8], [19]–[24],[29], but recognizing atypical facial expressions is a relatively unexplored area. This can be attributed to the lack of a sufficiently-large dataset to help researchers better understand such behaviors. There are current works that use images, but not video [10] but do not make use of crucial temporal information for behavior analysis with Transformer models [3], [13], [18] . Further works analyze behaviors via video [30], gaze analysis [1], [4], [12], [17], other behaviors [2], or via expensive equipment such as fMRI [6], [9], [11], [16], [26],[27], [31], which may differ in degree between neurodivergent individuals. By contrast, we analyze ASD-related behaviors by explicitly evoking reactions in more controlled settings.  \nData Collection. Each instance of our dataset contains a sequence of frame features for each video, thus obscuring the identify of the subjects, the ASD classification label (yes or no), and a text description. The frame features describe a person reacting to a sensory stimulus (e.g., taste or smell), with the video recording the subject’s reaction in terms of the head pose (behavior), and facial expression. The text descriptions provide a description of the expression for every six second interval, e.g., “The face expression changes from neutral to  \ndisgust” if the face expression changes within that time frame. Methodology. Our method consists of two main components. First is the encoder, which is itself composed to two subencoder modules. We have one model (EfficientNet B0 CNN [28]) to focus on neurotypical behaviors while the other model (ResNet-18 [7] pretrained for the facial expression recognition task on MS-Celeb [5]) focuses on more neurodivergent behaviors. These deep learning-based models focus on spatial features. The input consists of 32 frames for each video. The second main component is the decoder, which is implemented by a spatial transformer, e.g., a vision transformer [3] . We then use a fully-connected network (or MLP) to output the probabilites for the classes. When we sample video slices, we average the proba","cbCaij5hLswdzlBg","https://ap.wps.com/l/cbCaij5hLswdzlBg","pdf",121032,"English","# Abstract\n# Introduction and Related Work\n# Data Collection\n# Methodology\n# Experimental Results\n# References","[{\"question\":\"What problem does the Video ASD dataset target?\",\"answer\":\"It targets ASD-related behavior classification from video, capturing facial expressions and head-pose-linked actions under external stimuli, with special attention to neurodivergent variability.\"},{\"question\":\"How are video inputs represented in the dataset and model?\",\"answer\":\"Each dataset instance contains a sequence of frame features, alongside labels (yes/no) and timed text descriptions; the model processes 32 frames per video for spatial feature extraction and temporal decoding.\"},{\"question\":\"What experimental setup and performance results are reported?\",\"answer\":\"The framework is implemented in PyTorch and trained with 5-fold cross-validation using AdamW and cosine annealing, with face alignment and frame filtering; the reported test accuracy is 81.48% with an F1 score of 0.7289.\"}]","A Novel Dataset for Video-Based Neurodivergent Classification Leveraging Extra-Stimulatory Behavior | PDF"]