{"title":"Session details: Paper Session","authors":"S. Alletto","doi":"10.1145/3252801","DOIUrl":"https://doi.org/10.1145/3252801","url":null,"abstract":"","PeriodicalId":126678,"journal":{"name":"Proceedings of the 2017 Workshop on Wearable MultiMedia","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114904332","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}
We are delighted to welcome you to Workshop on Wearable Multimedia (WearMMe 2017) held in conjunction with ACM International Conference on Multimedia Retrieval (ICMR) in Bucharest, Romania. There has been substantial progress to date in developing computing devices and sensors that can be easily carried on the body. The last few years have also been marked by some notable achievements in learning from sensory data. This unique combination poses research challenges and opportunities for the next future of wearable computing. We believe wearable computing will be a very prominent research field for the multimedia and other communities. As such, there is a compelling need for science and technology that enable devices, algorithms, and humans to interact to achieve humanistic intelligence reciprocally. The range of real-world examples and applications of wearable is large and spans from the web and social applications (e.g. egocentric search engines, recommendation systems, and personalization), to medical robotics (e.g. assistive devices, bionic limbs and exoskeletons). The aim of this workshop is to bring together experts from various research communities including multimedia, computer vision, human-computer interaction, robotics, and machine learning to share recent advances and explore the future research. Toward this end, we are proud to have organized an exciting program in this half-day event. We are pleased to have Associate Professor Yusuke Sugano of Osaka University in Japan to give a keynote speech on appearance-based gaze estimation from ubiquitous cameras. We are also fortunate to have Dr. Kyriaki Kalimeri of ISI Foundation in Italy to share her recent work on identifying urban mobility challenges for the visually impaired with mobile monitoring of multimodal bio-signals. Last but not least, we are pleased to have three stimulating presentations selected from papers submitted to the workshop. Finally, we wish all the attendees a highly stimulating, informative, and enjoyable workshop.
{"title":"Proceedings of the 2017 Workshop on Wearable MultiMedia","authors":"S. Alletto, F. Pernici, Yoichi Sato","doi":"10.1145/3080538","DOIUrl":"https://doi.org/10.1145/3080538","url":null,"abstract":"We are delighted to welcome you to Workshop on Wearable Multimedia (WearMMe 2017) held in conjunction with ACM International Conference on Multimedia Retrieval (ICMR) in Bucharest, Romania. \u0000 \u0000There has been substantial progress to date in developing computing devices and sensors that can be easily carried on the body. The last few years have also been marked by some notable achievements in learning from sensory data. This unique combination poses research challenges and opportunities for the next future of wearable computing. We believe wearable computing will be a very prominent research field for the multimedia and other communities. As such, there is a compelling need for science and technology that enable devices, algorithms, and humans to interact to achieve humanistic intelligence reciprocally. The range of real-world examples and applications of wearable is large and spans from the web and social applications (e.g. egocentric search engines, recommendation systems, and personalization), to medical robotics (e.g. assistive devices, bionic limbs and exoskeletons). \u0000 \u0000The aim of this workshop is to bring together experts from various research communities including multimedia, computer vision, human-computer interaction, robotics, and machine learning to share recent advances and explore the future research. Toward this end, we are proud to have organized an exciting program in this half-day event. We are pleased to have Associate Professor Yusuke Sugano of Osaka University in Japan to give a keynote speech on appearance-based gaze estimation from ubiquitous cameras. We are also fortunate to have Dr. Kyriaki Kalimeri of ISI Foundation in Italy to share her recent work on identifying urban mobility challenges for the visually impaired with mobile monitoring of multimodal bio-signals. Last but not least, we are pleased to have three stimulating presentations selected from papers submitted to the workshop. \u0000 \u0000Finally, we wish all the attendees a highly stimulating, informative, and enjoyable workshop.","PeriodicalId":126678,"journal":{"name":"Proceedings of the 2017 Workshop on Wearable MultiMedia","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114790850","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}
Marco Godi, Maedeh Aghaei, Mariella Dimiccoli, M. Cristani
Clothing conveys a strong communicative message in terms of social signals, influencing the impression and behaviour of others towards a person; unfortunately, the nature of this message is not completely clear, and social signal processing approaches are starting to consider this problem. Wearable computing devices offer a unique perspective in this scenario, capturing fine details of clothing items in the same way we do during a social interaction, through ego-centered points of views. These clothing characteristics can be then employed to unveil statistical relations with personal impressions. This position paper investigates this novel research direction, individuating the main objectives, the possible problems, viable research strategies, techniques and expected results. This analysis gives birth to brand-new concepts such as clothing saliency, that is, those parts of garments more relevant for triggering personal impressions.
{"title":"Wearable for Wearable: A Social Signal Processing Perspective for Clothing Analysis using Wearable Devices","authors":"Marco Godi, Maedeh Aghaei, Mariella Dimiccoli, M. Cristani","doi":"10.1145/3080538.3080540","DOIUrl":"https://doi.org/10.1145/3080538.3080540","url":null,"abstract":"Clothing conveys a strong communicative message in terms of social signals, influencing the impression and behaviour of others towards a person; unfortunately, the nature of this message is not completely clear, and social signal processing approaches are starting to consider this problem. Wearable computing devices offer a unique perspective in this scenario, capturing fine details of clothing items in the same way we do during a social interaction, through ego-centered points of views. These clothing characteristics can be then employed to unveil statistical relations with personal impressions. This position paper investigates this novel research direction, individuating the main objectives, the possible problems, viable research strategies, techniques and expected results. This analysis gives birth to brand-new concepts such as clothing saliency, that is, those parts of garments more relevant for triggering personal impressions.","PeriodicalId":126678,"journal":{"name":"Proceedings of the 2017 Workshop on Wearable MultiMedia","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114344308","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}
Wearable cameras allow to easily acquire long and unstructured egocentric videos. In this context, temporal video segmentation methods can be useful to improve indexing, retrieval and summarization of such content. While past research investigated methods for temporal segmentation of egocentric videos according to different criteria (e.g., motion, location or appearance), many of them do not explicitly enforce any form of temporal coherence. Moreover, evaluations have been generally performed using frame-based measures, which only account for the overall correctness of predicted frames, overlooking the structure of the produced segmentation. In this paper, we investigate how a Hidden Markov Model based on an ad-hoc transition matrix can be exploited to obtain a more accurate segmentation from frame-based predictions in the context of location-based segmentation of egocentric videos. We introduce a segment-based evaluation measure which strongly penalizes over-segmented and under-segmented results. Experiments show that the exploitation of a Hidden Markov Model for temporal smoothing greatly improves temporal segmentation results and outperforms current video segmentation methods designed for both third-person and first-person videos.
{"title":"On the Exploitation of Hidden Markov Models to Improve Location-Based Temporal Segmentation of Egocentric Videos","authors":"Antonino Furnari, S. Battiato, G. Farinella","doi":"10.1145/3080538.3080539","DOIUrl":"https://doi.org/10.1145/3080538.3080539","url":null,"abstract":"Wearable cameras allow to easily acquire long and unstructured egocentric videos. In this context, temporal video segmentation methods can be useful to improve indexing, retrieval and summarization of such content. While past research investigated methods for temporal segmentation of egocentric videos according to different criteria (e.g., motion, location or appearance), many of them do not explicitly enforce any form of temporal coherence. Moreover, evaluations have been generally performed using frame-based measures, which only account for the overall correctness of predicted frames, overlooking the structure of the produced segmentation. In this paper, we investigate how a Hidden Markov Model based on an ad-hoc transition matrix can be exploited to obtain a more accurate segmentation from frame-based predictions in the context of location-based segmentation of egocentric videos. We introduce a segment-based evaluation measure which strongly penalizes over-segmented and under-segmented results. Experiments show that the exploitation of a Hidden Markov Model for temporal smoothing greatly improves temporal segmentation results and outperforms current video segmentation methods designed for both third-person and first-person videos.","PeriodicalId":126678,"journal":{"name":"Proceedings of the 2017 Workshop on Wearable MultiMedia","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128704798","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}
J. Benois-Pineau, M. García-Vázquez, L. Moralez, A. A. Ramírez-Acosta
Recognition of objects of a given category in visual content is one of the key problems in computer vision and multimedia. It is strongly needed in wearable video shooting for a wide range of important applications in society. Supervised learning approaches are proved to be the most efficient in this task. They require available ground truth for training models. It is specifically true for Deep Convolution Networks, but is also hold for other popular models such as SVM on visual signatures. Annotation of ground truth when drawing bounding boxes (BB) is a very tedious task requiring important human resource. The research in prediction of visual attention in images and videos has attained maturity, specifically in what concerns bottom-up visual attention modeling. Hence, instead of annotating the ground truth manually with BB we propose to use automatically predicted salient areas as object locators for annotation. Such a prediction of saliency is not perfect, nevertheless. Hence active contours models on saliency maps are used in order to isolate the most prominent areas covering the objects. The approach is tested in the framework of a well-studied supervised learning model by SVM with psycho-visual weighted Bag-of-Words. An egocentric GTEA dataset was used in the experiment. The difference in mAP (mean average precision) is less than 10 percent while the mean annotation time is 36% lower.
{"title":"Semi-Automatic Annotation with Predicted Visual Saliency Maps for Object Recognition in Wearable Video","authors":"J. Benois-Pineau, M. García-Vázquez, L. Moralez, A. A. Ramírez-Acosta","doi":"10.1145/3080538.3080541","DOIUrl":"https://doi.org/10.1145/3080538.3080541","url":null,"abstract":"Recognition of objects of a given category in visual content is one of the key problems in computer vision and multimedia. It is strongly needed in wearable video shooting for a wide range of important applications in society. Supervised learning approaches are proved to be the most efficient in this task. They require available ground truth for training models. It is specifically true for Deep Convolution Networks, but is also hold for other popular models such as SVM on visual signatures. Annotation of ground truth when drawing bounding boxes (BB) is a very tedious task requiring important human resource. The research in prediction of visual attention in images and videos has attained maturity, specifically in what concerns bottom-up visual attention modeling. Hence, instead of annotating the ground truth manually with BB we propose to use automatically predicted salient areas as object locators for annotation. Such a prediction of saliency is not perfect, nevertheless. Hence active contours models on saliency maps are used in order to isolate the most prominent areas covering the objects. The approach is tested in the framework of a well-studied supervised learning model by SVM with psycho-visual weighted Bag-of-Words. An egocentric GTEA dataset was used in the experiment. The difference in mAP (mean average precision) is less than 10 percent while the mean annotation time is 36% lower.","PeriodicalId":126678,"journal":{"name":"Proceedings of the 2017 Workshop on Wearable MultiMedia","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114409905","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：