{"title":"Session details: Session 1: Visual Recognition of Families In the Wild: A Big Data Challenge","authors":"Timothy Gillis","doi":"10.1145/3247931","DOIUrl":"https://doi.org/10.1145/3247931","url":null,"abstract":"","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131697033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Joseph P. Robinson, Ming Shao, Handong Zhao, Yue Wu, Timothy Gillis, Y. Fu
Recognizing Families In the Wild (RFIW) is a large-scale, multi-track automatic kinship recognition evaluation, supporting both kinship verification and family classification on scales much larger than ever before. It was organized as a Data Challenge Workshop hosted in conjunction with ACM Multimedia 2017. This was achieved with the largest image collection that supports kin-based vision tasks. In the end, we use this manuscript to summarize evaluation protocols, progress made and some technical background and performance ratings of the algorithms used, and a discussion on promising directions for both research and engineers to be taken next in this line of work.
{"title":"Recognizing Families In the Wild (RFIW): Data Challenge Workshop in conjunction with ACM MM 2017","authors":"Joseph P. Robinson, Ming Shao, Handong Zhao, Yue Wu, Timothy Gillis, Y. Fu","doi":"10.1145/3134421.3134424","DOIUrl":"https://doi.org/10.1145/3134421.3134424","url":null,"abstract":"Recognizing Families In the Wild (RFIW) is a large-scale, multi-track automatic kinship recognition evaluation, supporting both kinship verification and family classification on scales much larger than ever before. It was organized as a Data Challenge Workshop hosted in conjunction with ACM Multimedia 2017. This was achieved with the largest image collection that supports kin-based vision tasks. In the end, we use this manuscript to summarize evaluation protocols, progress made and some technical background and performance ratings of the algorithms used, and a discussion on promising directions for both research and engineers to be taken next in this line of work.","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133075645","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
It is our great pleasure to welcome you to the 2017 Recognizing Families In the Wild (RFIW) -- RFIW'17, based on the 1st ever large-scale visual kinship recognition data challenge workshop. RFIW'17 was organized with Families In the Wild (FIW)-the largest image database for kinship recognition. Kinship recognition has an abundance of practical uses. Including, but not limited to, forensic investigations, photo library management, historic lineage & genealogical studies, social-media- based analysis, cases of missing children & human tracking, and problems of immigration & border patrol. This workshop will cover the results of the data challenge, the current status and plans for FIW, and the past, present and future of kinship recognition. RFIW gives researchers and practitioners a unique opportunity to exchange their perspectives with others interested in the various aspects of kinship recognition technologies. Participants of the competition spanned all over the globe. We had nearly 100 teams register for the event, with 12 teams submitting competitive results, and 6 of which had their work published. RFIW'17 workshop is on 27 October 2017 at ACM MM. The agenda includes the following: RFIW Data Workshop Introduction and Overview Oral presentations (2) given by top performers of RFIW Poster Presentations by all other authors during coffee break Past, Present, and Future of Kinship Recognition, FIW, and upcoming events Additionally, we have two great keynotes coming to share complimentary views of the problem: Rapid DNA Performance Results on Family Relationship Verification, Christopher Miles (Department of Homeland Security) Recent Advances in Deep Reinforcement Learning for Computer Vision and NLP, Caiming Xiong (Salesforce Research) We encourage attendees to join us to learn from these valuable and insightful speakers that will guide us to a better understanding of the future.
{"title":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","authors":"Y. Fu, Joseph P. Robinson, Ming Shao, Siyu Xia","doi":"10.1145/3134421","DOIUrl":"https://doi.org/10.1145/3134421","url":null,"abstract":"It is our great pleasure to welcome you to the 2017 Recognizing Families In the Wild (RFIW) -- RFIW'17, based on the 1st ever large-scale visual kinship recognition data challenge workshop. RFIW'17 was organized with Families In the Wild (FIW)-the largest image database for kinship recognition. Kinship recognition has an abundance of practical uses. Including, but not limited to, forensic investigations, photo library management, historic lineage & genealogical studies, social-media- based analysis, cases of missing children & human tracking, and problems of immigration & border patrol. This workshop will cover the results of the data challenge, the current status and plans for FIW, and the past, present and future of kinship recognition. RFIW gives researchers and practitioners a unique opportunity to exchange their perspectives with others interested in the various aspects of kinship recognition technologies. \u0000 \u0000Participants of the competition spanned all over the globe. We had nearly 100 teams register for the event, with 12 teams submitting competitive results, and 6 of which had their work published. \u0000 \u0000RFIW'17 workshop is on 27 October 2017 at ACM MM. The agenda includes the following: \u0000RFIW Data Workshop Introduction and Overview \u0000Oral presentations (2) given by top performers of RFIW \u0000Poster Presentations by all other authors during coffee break \u0000Past, Present, and Future of Kinship Recognition, FIW, and upcoming events \u0000 \u0000 \u0000 \u0000Additionally, we have two great keynotes coming to share complimentary views of the problem: \u0000Rapid DNA Performance Results on Family Relationship Verification, Christopher Miles (Department of Homeland Security) \u0000Recent Advances in Deep Reinforcement Learning for Computer Vision and NLP, Caiming Xiong (Salesforce Research) \u0000 \u0000 \u0000 \u0000We encourage attendees to join us to learn from these valuable and insightful speakers that will guide us to a better understanding of the future.","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130568752","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kinship Verification of two or more people has shown to be a complicated problem, though it is widely used in various practical tasks and applications. The areas of the use-cases vary. Among them are applications for homeland security, automatic family recognition, youth and elder matching or predicting and more. We propose using Deep Learning approach to deal with the problem of Kin Verification, such to provide a logical explanation for solving the problem with a novel mechanism for training on the FIW data-set. Our method obtains state-of-the-art for the FIW challenge for the restricted-image setting11
{"title":"Kin-Verification Model on FIW Dataset Using Multi-Set Learning and Local Features","authors":"Eran Dahan, Y. Keller, Shahar Mahpod","doi":"10.1145/3134421.3134423","DOIUrl":"https://doi.org/10.1145/3134421.3134423","url":null,"abstract":"Kinship Verification of two or more people has shown to be a complicated problem, though it is widely used in various practical tasks and applications. The areas of the use-cases vary. Among them are applications for homeland security, automatic family recognition, youth and elder matching or predicting and more. We propose using Deep Learning approach to deal with the problem of Kin Verification, such to provide a logical explanation for solving the problem with a novel mechanism for training on the FIW data-set. Our method obtains state-of-the-art for the FIW challenge for the restricted-image setting11","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114542295","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Session details: Keynote & Invited Talks","authors":"Joseph P. Robinson","doi":"10.1145/3247930","DOIUrl":"https://doi.org/10.1145/3247930","url":null,"abstract":"","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122094139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Deep reinforcement learning is considered as a way of building autonomous system with a higher level understanding of the world and would revolutionize the field of AI. Recently, some researchers have made many progresses such as learning to play video games like Atari, learning control policy for robots from camera input. In this talk, we begin with general introduction of deep reinforcement learning algorithms, including policy optimization, deep Qlearning, then we will highlight the progresses that have achieved in Vision and NLP via DRL.
{"title":"Recent Progress in Deep Reinforcement Learning for Computer Vision and NLP","authors":"Caiming Xiong","doi":"10.1145/3134421.3137039","DOIUrl":"https://doi.org/10.1145/3134421.3137039","url":null,"abstract":"Deep reinforcement learning is considered as a way of building autonomous system with a higher level understanding of the world and would revolutionize the field of AI. Recently, some researchers have made many progresses such as learning to play video games like Atari, learning control policy for robots from camera input. In this talk, we begin with general introduction of deep reinforcement learning algorithms, including policy optimization, deep Qlearning, then we will highlight the progresses that have achieved in Vision and NLP via DRL.","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125996145","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Yong Li, Jiabei Zeng, Jie Zhang, Anbo Dai, Meina Kan, S. Shan, Xilin Chen
Automatic kinship verification has attracted increasing attentions as it holds promise to an abundance of applications. However, existing kinship verification methods suffer from the lack of large scale real-world data. Without enough training data, it is difficult to learn proper features that are discriminant for blood-related peoples. In this work, we propose KinNet, a fine-to-coarse deep metric learning framework for kinship verification. In the framework, we transfer knowledge from the large-scale-data-driven face recognition task, which is a fine-grained version of kinship recognition, by pre-training the network with massive data for face recognition. Then, the network is fine-tuned to find a metric space where kin-related peoples are discriminant. The metric space is learned by minimizing a soft triplet loss on the augmented kinship dataset. An augmented strategy is proposed to balance the amount of images per family member. Finally, we ensemble four networks to further boost the performance. The experimental results on the 1st Large-Scale Kinship Recognition Data Challenge (Track 1) demonstrate that our KinNet achieves the state-of-the-art performance in kinship verification.
{"title":"KinNet","authors":"Yong Li, Jiabei Zeng, Jie Zhang, Anbo Dai, Meina Kan, S. Shan, Xilin Chen","doi":"10.1145/3134421.3134425","DOIUrl":"https://doi.org/10.1145/3134421.3134425","url":null,"abstract":"Automatic kinship verification has attracted increasing attentions as it holds promise to an abundance of applications. However, existing kinship verification methods suffer from the lack of large scale real-world data. Without enough training data, it is difficult to learn proper features that are discriminant for blood-related peoples. In this work, we propose KinNet, a fine-to-coarse deep metric learning framework for kinship verification. In the framework, we transfer knowledge from the large-scale-data-driven face recognition task, which is a fine-grained version of kinship recognition, by pre-training the network with massive data for face recognition. Then, the network is fine-tuned to find a metric space where kin-related peoples are discriminant. The metric space is learned by minimizing a soft triplet loss on the augmented kinship dataset. An augmented strategy is proposed to balance the amount of images per family member. Finally, we ensemble four networks to further boost the performance. The experimental results on the 1st Large-Scale Kinship Recognition Data Challenge (Track 1) demonstrate that our KinNet achieves the state-of-the-art performance in kinship verification.","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127106060","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Kinship verification in the wild is an interesting and challenging problem. The goal of kinship verification is to determine whether a pair of faces are blood relatives or not. Most previous methods for kinship verification can be divided as hand-crafted features based shallow learning methods and convolutional neural network (CNN) based deep learning methods. Nevertheless, these methods are still posed with the challenging task of recognizing kinship cues from facial images. Part of the reason for this may be that, the family information and the distribution difference of pairwise kin-face data based kinship verification issue are rarely considered. Inspired by maximum mean discrepancy (MMD) and generative adversarial net (GAN), family ID based Adversarial contrastive residual Network (AdvNet) is proposed for large-scale (1 Million) kinship recognition in this paper. The MMD based adversarial loss (AL), pairwise contrastive loss (CL) and family ID based softmax loss (SL) are jointly formulated in the proposed AdvNet for kin-relation enhancement and discovery. Further, the deep nets ensemble is used for deep kin-feature augmentation. Finally, Euclidean distance metric is used for kinship recognition. Extensive experiments on the 1st Large-Scale Kinship Recognition Data Challenge (Families in the wild) show the effectiveness of our proposed AdvNet and ensemble based feature augmentation.
{"title":"AdvNet: Adversarial Contrastive Residual Net for 1 Million Kinship Recognition","authors":"Qingyan Duan, Lei Zhang","doi":"10.1145/3134421.3134422","DOIUrl":"https://doi.org/10.1145/3134421.3134422","url":null,"abstract":"Kinship verification in the wild is an interesting and challenging problem. The goal of kinship verification is to determine whether a pair of faces are blood relatives or not. Most previous methods for kinship verification can be divided as hand-crafted features based shallow learning methods and convolutional neural network (CNN) based deep learning methods. Nevertheless, these methods are still posed with the challenging task of recognizing kinship cues from facial images. Part of the reason for this may be that, the family information and the distribution difference of pairwise kin-face data based kinship verification issue are rarely considered. Inspired by maximum mean discrepancy (MMD) and generative adversarial net (GAN), family ID based Adversarial contrastive residual Network (AdvNet) is proposed for large-scale (1 Million) kinship recognition in this paper. The MMD based adversarial loss (AL), pairwise contrastive loss (CL) and family ID based softmax loss (SL) are jointly formulated in the proposed AdvNet for kin-relation enhancement and discovery. Further, the deep nets ensemble is used for deep kin-feature augmentation. Finally, Euclidean distance metric is used for kinship recognition. Extensive experiments on the 1st Large-Scale Kinship Recognition Data Challenge (Families in the wild) show the effectiveness of our proposed AdvNet and ensemble based feature augmentation.","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127489078","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Oualid Laiadi, A. Ouamane, A. Benakcha, A. Taleb-Ahmed
As a part of the RFIW 2017 Data Challenge Workshop, we demonstrate performance on the large-scale FIW dataset, along with several pre-existing image collections. Noticing available version of SIEDA method work well on smaller datasets (i.e. Cornell and UB KinFace datasets) than on FIW, we propose modifications to address this disparity in results.
{"title":"RFIW 2017: LPQ-SIEDA for Large Scale Kinship Verification","authors":"Oualid Laiadi, A. Ouamane, A. Benakcha, A. Taleb-Ahmed","doi":"10.1145/3134421.3134426","DOIUrl":"https://doi.org/10.1145/3134421.3134426","url":null,"abstract":"As a part of the RFIW 2017 Data Challenge Workshop, we demonstrate performance on the large-scale FIW dataset, along with several pre-existing image collections. Noticing available version of SIEDA method work well on smaller datasets (i.e. Cornell and UB KinFace datasets) than on FIW, we propose modifications to address this disparity in results.","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116795245","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The Department of Homeland Security (DHS) Science and Technology Directorate (S&T), working with other Federal partners, developed Rapid DNA technology to meet a DHS requirement to verify family relationship claims to reduce fraud and expedite legal immigration. Several hundred thousand immigration relationship tests are processed annually for DHS by AABB accredited laboratories. The AABB Relationship Testing Subcommittee sets the standards for those tests and accreditation of relationship testing laboratories. But, shipping of the collection kits overseas and coordinating DNA collection can cause these tests to take weeks to months to process and results to be returned. Rapid DNA uses microfluidics technology to reduce the million-dollar clean-room DNA processes down to a standalone, integrated, and automated desktop system that processes five to seven DNA samples in 90 minutes in a closed, hands-off system. Rapid DNA is built to be operated by DHS field officers and agents. It is about the size of a laser printer, and can be operated in an office setting by anyone who can change a printer cartridge. It is also ruggedized to military requirements, and can be operated in hazardous field environments on generator power. Setup takes just 15 minutes. With success showing that Rapid DNA accurately verifies family relationships, additional DHS needs were identified to fight human trafficking along U.S. borders and reunify families following critical or mass-casualty incidents. This presentation will discuss the DHS needs for family relationship verification, AABB Standards, and Rapid DNA performance and field test results.
{"title":"Rapid DNA Performance Results on Family Relationship Verification","authors":"Christopher Miles","doi":"10.1145/3134421.3137040","DOIUrl":"https://doi.org/10.1145/3134421.3137040","url":null,"abstract":"The Department of Homeland Security (DHS) Science and Technology Directorate (S&T), working with other Federal partners, developed Rapid DNA technology to meet a DHS requirement to verify family relationship claims to reduce fraud and expedite legal immigration. Several hundred thousand immigration relationship tests are processed annually for DHS by AABB accredited laboratories. The AABB Relationship Testing Subcommittee sets the standards for those tests and accreditation of relationship testing laboratories. But, shipping of the collection kits overseas and coordinating DNA collection can cause these tests to take weeks to months to process and results to be returned. Rapid DNA uses microfluidics technology to reduce the million-dollar clean-room DNA processes down to a standalone, integrated, and automated desktop system that processes five to seven DNA samples in 90 minutes in a closed, hands-off system. Rapid DNA is built to be operated by DHS field officers and agents. It is about the size of a laser printer, and can be operated in an office setting by anyone who can change a printer cartridge. It is also ruggedized to military requirements, and can be operated in hazardous field environments on generator power. Setup takes just 15 minutes. With success showing that Rapid DNA accurately verifies family relationships, additional DHS needs were identified to fight human trafficking along U.S. borders and reunify families following critical or mass-casualty incidents. This presentation will discuss the DHS needs for family relationship verification, AABB Standards, and Rapid DNA performance and field test results.","PeriodicalId":209776,"journal":{"name":"Proceedings of the 2017 Workshop on Recognizing Families In the Wild","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2017-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129486421","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：