Marta Kipke, L. Brinkmeyer, Souaybou Bagayoko, L. Schmidt-Thieme, Martin Langner
{"title":"基于对象检测的希腊花瓶画家归属深度标注","authors":"Marta Kipke, L. Brinkmeyer, Souaybou Bagayoko, L. Schmidt-Thieme, Martin Langner","doi":"10.1145/3552464.3555684","DOIUrl":null,"url":null,"abstract":"Painter attribution is based on a variety of factors, oftentimes deeply buried in the details such as the brushstrokes of the ears or the eyes, which a painter might paint in a specific way. To get to this details, the images have to be examined carefully and intensively. Our work is focused on this phenomenon of painter attribution, investigating those details using supervised machine learning methods for image recognition that rely on a set representation. In this paper however, we are going to focus on one step of our work specifically: The annotation process. With such a focus on details, a dense and detailed, but also transparent annotation of the images is necessary. On one hand this is essential for our research, on the other hand however, it is very time consuming and requires a lot of human resources. Therefore we developed an ontology for the annotation of the images and a semi-automated workflow with object detection component using YOLOv3 and closely tied to our ontology. This way we were able to automate our processes as efficiently as possible while maintaining the complexity of our annotations.","PeriodicalId":131418,"journal":{"name":"Proceedings of the 4th ACM International workshop on Structuring and Understanding of Multimedia heritAge Contents","volume":"10 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Deep Level Annotation for Painter Attribution on Greek Vases utilizing Object Detection\",\"authors\":\"Marta Kipke, L. Brinkmeyer, Souaybou Bagayoko, L. Schmidt-Thieme, Martin Langner\",\"doi\":\"10.1145/3552464.3555684\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Painter attribution is based on a variety of factors, oftentimes deeply buried in the details such as the brushstrokes of the ears or the eyes, which a painter might paint in a specific way. To get to this details, the images have to be examined carefully and intensively. Our work is focused on this phenomenon of painter attribution, investigating those details using supervised machine learning methods for image recognition that rely on a set representation. In this paper however, we are going to focus on one step of our work specifically: The annotation process. With such a focus on details, a dense and detailed, but also transparent annotation of the images is necessary. On one hand this is essential for our research, on the other hand however, it is very time consuming and requires a lot of human resources. Therefore we developed an ontology for the annotation of the images and a semi-automated workflow with object detection component using YOLOv3 and closely tied to our ontology. This way we were able to automate our processes as efficiently as possible while maintaining the complexity of our annotations.\",\"PeriodicalId\":131418,\"journal\":{\"name\":\"Proceedings of the 4th ACM International workshop on Structuring and Understanding of Multimedia heritAge Contents\",\"volume\":\"10 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-10-10\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 4th ACM International workshop on Structuring and Understanding of Multimedia heritAge Contents\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/3552464.3555684\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 4th ACM International workshop on Structuring and Understanding of Multimedia heritAge Contents","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3552464.3555684","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Deep Level Annotation for Painter Attribution on Greek Vases utilizing Object Detection
Painter attribution is based on a variety of factors, oftentimes deeply buried in the details such as the brushstrokes of the ears or the eyes, which a painter might paint in a specific way. To get to this details, the images have to be examined carefully and intensively. Our work is focused on this phenomenon of painter attribution, investigating those details using supervised machine learning methods for image recognition that rely on a set representation. In this paper however, we are going to focus on one step of our work specifically: The annotation process. With such a focus on details, a dense and detailed, but also transparent annotation of the images is necessary. On one hand this is essential for our research, on the other hand however, it is very time consuming and requires a lot of human resources. Therefore we developed an ontology for the annotation of the images and a semi-automated workflow with object detection component using YOLOv3 and closely tied to our ontology. This way we were able to automate our processes as efficiently as possible while maintaining the complexity of our annotations.