[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81743-en":3,"doc-seo-81743-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},81743,4398048949847,"Eliana","https://ap-avatar.wpscdn.com/avatar/400002536579ef2da7f?_k=1778318612642679267",8,"Research & Report","Enhancing Oracle Bone Inscription Recognition via Multi-Scale Layer Attention","Oracle Bone Inscriptions (OBIs) recognition is critical for interpreting ancient Chinese culture, yet remains difficult because OBIs exhibit complex, irregular, and degraded shapes. Prior approaches depend on expert knowledge and manual analysis, while deep learning methods often fail to capture the fine-grained details and subtle variations required for strong performance. Even state-of-the-art layer attention yields only marginal gains in OBI recognition. This work proposes Multi-Scale Layer Attention (MSLA) to jointly model multi-scale and cross-layer interactions, improving accuracy and robustness while preserving computational efficiency on large datasets.","Enhancing Oracle Bone Inscription Recognition via Multi-Scale Layer Attention  \nChaowen Yana , Kaishen Wangb , Yong Wanga , Jianlong Xiongb and Tao Hec,∗  \na Academy of Chinese Traditional Culture, Sichuan Normal University, Chengdu, China b College of Computer Science, Sichuan University, Chengdu, China  \nc School of Artificial Intelligence, Sichuan University, Chengdu, China  \narXiv :2607 .00057v1 [ cs .CV] 30 Jun 2026  \nARTICLE INFO  \nKeywords:  \nOracle Bone Inscriptions Layer Attention Mechanism Multi-Scale Recognition  \nAB STRACT  \nOracle Bone Inscriptions (OBIs) recognition plays a crucial role in understanding ancient Chinese culture. However, accurately recognizing OBIs remains highly challenging due to their complex, irregular, and often degraded shapes. Traditional methods rely on expert knowledge and manual analysis, which are time-consuming and error-prone. Although deep learning has greatly advanced general image recognition, existing methods struggle to capture the fine-grained details and subtle variations inherent in OBIs, resulting in limited performance. Even most recent and effective layer attention techniques are designed to capture fine-grained dependencies through enhanced inter-layer interactions, yet they still exhibit only marginal improvements in OBIs recognition. To address these limitations, we propose Multi-Scale Layer Attention (MSLA), a novel paradigm that explicitly models both multi-scale and cross-layer feature interactions. By enriching the representation with finegrained details across multiple spatial scales, MSLA enables more accurate and robust OBIs recognition. Extensive experiments on large-scale OBIs datasets demonstrate that MSLA consistently outperforms existing attention mechanisms while maintaining computational efficiency.  \n1. Introduction  \nOracle Bone Inscriptions (OBIs) [14, 28, 49], shown in Figure 1(a), represent the earliest known systematized form of Chinese writing and carry profound historical and cultural significance, forming the foundation of modern Chinese characters. Accurate recognition of OBIs is essential not only for deciphering ancient historical records but also for advancing our understanding of the origins and evolution of Chinese civilization [30, 3] . However, the complex, irregular, and often fragmented structures of OBIs make automatic recognition particularly challenging.  \nTraditionally, the recognition of OBIs relied on expert knowledge and manual analysis [30, 54, 52, 47], which is labor-intensive and prone to errors. With the advent of largescale OBI datasets [25, 61, 58, 57], deep learning techniques have been leveraged to improve recognition efficiency and accuracy. Architectures such as AlexNet [31], ResNet [19], and ViT [12] have demonstrated remarkable performance in conventional image recognition tasks, offering potential for addressing the complex and diverse characteristics of OBIs through hierarchical feature learning.  \nTo further enhance the model’s representational capacity and task performance, attention mechanisms [55] have been incorporated into image recognition tasks. These approaches—such as channel attention [23, 64, 62], spatial attention [63, 5], branch attention [51, 36], and temporal attention [67, 6]—aim to make feature extraction more effective by selectively emphasizing the most informative channels, spatial regions, branches, or temporal components, thereby enhancing recognition performance.  \n [tao_he@scu.edu.cn](tao_he@scu.edu.cn) (T. He) ORCID(s): 0000-0001-9405-3979 (T. He)  \nAlthough effective in general-purpose domains such as ImageNet [11], Cityscapes [7], and COCO [37], these methods exhibit limited performance when transferred to the specialized task of OBIs recognition. For instance, when applied to the large-scale OBIs dataset HUST-OBS [61] using ResNet-50 as the backbone, the aforementioned methods achieved comparable performance but introduced additional parameters and increased training costs. We attribute th","cbCaitiskNzNBokt","https://ap.wps.com/l/cbCaitiskNzNBokt","pdf",11698921,4,1,13,"English","en",105,"# Introduction\n## Background and Challenges of OBI Recognition\n## Attention Mechanisms in Image Recognition\n## Layer Attention and Its Limitations for OBIs\n## Proposed Solution: Multi-Scale Layer Attention (MSLA)","[{\"question\":\"Why is Oracle Bone Inscriptions recognition particularly challenging for automatic methods?\",\"answer\":\"OBIs have complex, irregular, and often fragmented or degraded shapes, which makes modeling fine-grained details and subtle variations difficult.\"},{\"question\":\"What limitation is observed in existing layer attention techniques for OBIs recognition?\",\"answer\":\"Although layer attention improves inter-layer interactions, it still delivers only marginal improvements for OBIs, indicating it cannot capture the required fine-grained dependencies well enough.\"},{\"question\":\"How does Multi-Scale Layer Attention (MSLA) address these limitations?\",\"answer\":\"MSLA explicitly models both multi-scale and cross-layer feature interactions, enriching representations with fine-grained details across multiple spatial scales to achieve more accurate and robust recognition.\"}]",1784175788,33,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"enhancing-oracle-bone-inscription-recognition-via-multi-scale-layer-attention","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":20},"https://docshare.wps.com/document/enhancing-oracle-bone-inscription-recognition-via-multi-scale-layer-attention/81743/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-26","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why is Oracle Bone Inscriptions recognition particularly challenging for automatic methods?","Question",{"text":75,"@type":76},"OBIs have complex, irregular, and often fragmented or degraded shapes, which makes modeling fine-grained details and subtle variations difficult.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What limitation is observed in existing layer attention techniques for OBIs recognition?",{"text":80,"@type":76},"Although layer attention improves inter-layer interactions, it still delivers only marginal improvements for OBIs, indicating it cannot capture the required fine-grained dependencies well enough.",{"name":82,"@type":73,"acceptedAnswer":83},"How does Multi-Scale Layer Attention (MSLA) address these limitations?",{"text":84,"@type":76},"MSLA explicitly models both multi-scale and cross-layer feature interactions, enriching representations with fine-grained details across multiple spatial scales to achieve more accurate and robust recognition.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]