WEKO3
アイテム
{"_buckets": {"deposit": "8e1bd9db-bfa1-4966-8a3c-93a7325c3b2c"}, "_deposit": {"id": "4857", "owners": [], "pid": {"revision_id": 0, "type": "depid", "value": "4857"}, "status": "published"}, "_oai": {"id": "oai:repository.dl.itc.u-tokyo.ac.jp:00004857", "sets": ["234", "262"]}, "item_7_alternative_title_1": {"attribute_name": "その他のタイトル", "attribute_value_mlt": [{"subitem_alternative_title": "皮膚色素モデルに基づく近赤外顔画像から可視光顔画像への変換"}]}, "item_7_biblio_info_7": {"attribute_name": "書誌情報", "attribute_value_mlt": [{"bibliographicIssueDates": {"bibliographicIssueDate": "2013-03-25", "bibliographicIssueDateType": "Issued"}, "bibliographic_titles": [{}]}]}, "item_7_date_granted_25": {"attribute_name": "学位授与年月日", "attribute_value_mlt": [{"subitem_dategranted": "2013-03-25"}]}, "item_7_degree_name_20": {"attribute_name": "学位名", "attribute_value_mlt": [{"subitem_degreename": "修士(情報理工学)"}]}, "item_7_description_5": {"attribute_name": "抄録", "attribute_value_mlt": [{"subitem_description": "Facial images are the primary biometric used for a human identification system and a tracking system. Especially, a visible light (VIS) facial image is commonly used as it is easy to be acquired and its appearance is similar to the human visual system. However, images taken under varying illumination or a low-lit condition result in poor recognition accuracy for systems using VIS images. To alleviate this problem, a system using an active near infrared (NIR) flash is proposed to acquire illumination invariant facial images. Yet, this system requires both probe and gallery images to be taken under NIR spectrum which limits the applicability of the system as NIR images are often unavailable for gallery images, e.g., photos on a passport or a driving license. In such cases, probe NIR images are matched against VIS gallery images. One of the solutions to this problem is to convert probe NIR images to artificial VIS images before matching. In this thesis, we propose a physics-based method to synthesize artificial facial images in visible wavelengths from multiband NIR images. Studies on photometric properties of human skin show that melanin and hemoglobin components are dominant factors that affect the skin appearance under different light spectrum. Our physics-based synthesis method takes account of the photometric properties of human skin, introducing a skin pigment model. Specifically, a set of intensities observed at a certain surface point with varying wavelength is represented by a linear combination of both the pigment components. Our method learns the spectral basis vectors, which describe absorbance of both the pigments, from multispectral image dataset by using Independent Component Analysis (ICA). Then, our method estimates the coefficients, which are pixel-wise densities of both the pigments, from a multiband NIR image, and finally converts it to a visible light image. We demonstrate that our method works well for real facial images even though only a small dataset is available for learning basis vectors.", "subitem_description_type": "Abstract"}]}, "item_7_full_name_3": {"attribute_name": "著者別名", "attribute_value_mlt": [{"nameIdentifiers": [{"nameIdentifier": "10554", "nameIdentifierScheme": "WEKO"}], "names": [{"name": "ゴウ, キム シン"}]}]}, "item_7_select_21": {"attribute_name": "学位", "attribute_value_mlt": [{"subitem_select_item": "master"}]}, "item_7_text_24": {"attribute_name": "研究科・専攻", "attribute_value_mlt": [{"subitem_text_value": "情報理工学系研究科・電子情報学専攻"}]}, "item_7_text_36": {"attribute_name": "資源タイプ", "attribute_value_mlt": [{"subitem_text_value": "Thesis"}]}, "item_7_text_4": {"attribute_name": "著者所属", "attribute_value_mlt": [{"subitem_text_value": "東京大学大学院情報理工学系研究科電子情報学専攻"}, {"subitem_text_value": "Department of Information and Communication Engineering, Graduate School of Information Science and Technology, The University of Tokyo"}]}, "item_creator": {"attribute_name": "著者", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "Goh, Kim Shing"}], "nameIdentifiers": [{"nameIdentifier": "10553", "nameIdentifierScheme": "WEKO"}]}]}, "item_files": {"attribute_name": "ファイル情報", "attribute_type": "file", "attribute_value_mlt": [{"accessrole": "open_date", "date": [{"dateType": "Available", "dateValue": "2017-06-01"}], "displaytype": "detail", "download_preview_message": "", "file_order": 0, "filename": "48116421.pdf", "filesize": [{"value": "9.4 MB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensetype": "license_free", "mimetype": "application/pdf", "size": 9400000.0, "url": {"label": "48116421.pdf", "url": "https://repository.dl.itc.u-tokyo.ac.jp/record/4857/files/48116421.pdf"}, "version_id": "8752de1a-5fa1-4e32-b6f1-a5d3c2d42c48"}]}, "item_language": {"attribute_name": "言語", "attribute_value_mlt": [{"subitem_language": "eng"}]}, "item_resource_type": {"attribute_name": "資源タイプ", "attribute_value_mlt": [{"resourcetype": "thesis", "resourceuri": "http://purl.org/coar/resource_type/c_46ec"}]}, "item_title": "Converting Near Infrared Facial Images to Visible Light Images using Skin Pigment Model", "item_titles": {"attribute_name": "タイトル", "attribute_value_mlt": [{"subitem_title": "Converting Near Infrared Facial Images to Visible Light Images using Skin Pigment Model"}]}, "item_type_id": "7", "owner": "1", "path": ["234", "262"], "permalink_uri": "http://hdl.handle.net/2261/54202", "pubdate": {"attribute_name": "公開日", "attribute_value": "2013-05-07"}, "publish_date": "2013-05-07", "publish_status": "0", "recid": "4857", "relation": {}, "relation_version_is_last": true, "title": ["Converting Near Infrared Facial Images to Visible Light Images using Skin Pigment Model"], "weko_shared_id": null}
Converting Near Infrared Facial Images to Visible Light Images using Skin Pigment Model
http://hdl.handle.net/2261/54202
http://hdl.handle.net/2261/5420295b6b25f-fbbf-4c23-bcb2-b5de1844cbc6
名前 / ファイル | ライセンス | アクション |
---|---|---|
48116421.pdf (9.4 MB)
|
|
Item type | 学位論文 / Thesis or Dissertation(1) | |||||
---|---|---|---|---|---|---|
公開日 | 2013-05-07 | |||||
タイトル | ||||||
タイトル | Converting Near Infrared Facial Images to Visible Light Images using Skin Pigment Model | |||||
言語 | ||||||
言語 | eng | |||||
資源タイプ | ||||||
資源 | http://purl.org/coar/resource_type/c_46ec | |||||
タイプ | thesis | |||||
その他のタイトル | ||||||
その他のタイトル | 皮膚色素モデルに基づく近赤外顔画像から可視光顔画像への変換 | |||||
著者 |
Goh, Kim Shing
× Goh, Kim Shing |
|||||
著者別名 | ||||||
識別子 | 10554 | |||||
識別子Scheme | WEKO | |||||
姓名 | ゴウ, キム シン | |||||
著者所属 | ||||||
著者所属 | 東京大学大学院情報理工学系研究科電子情報学専攻 | |||||
著者所属 | ||||||
著者所属 | Department of Information and Communication Engineering, Graduate School of Information Science and Technology, The University of Tokyo | |||||
Abstract | ||||||
内容記述タイプ | Abstract | |||||
内容記述 | Facial images are the primary biometric used for a human identification system and a tracking system. Especially, a visible light (VIS) facial image is commonly used as it is easy to be acquired and its appearance is similar to the human visual system. However, images taken under varying illumination or a low-lit condition result in poor recognition accuracy for systems using VIS images. To alleviate this problem, a system using an active near infrared (NIR) flash is proposed to acquire illumination invariant facial images. Yet, this system requires both probe and gallery images to be taken under NIR spectrum which limits the applicability of the system as NIR images are often unavailable for gallery images, e.g., photos on a passport or a driving license. In such cases, probe NIR images are matched against VIS gallery images. One of the solutions to this problem is to convert probe NIR images to artificial VIS images before matching. In this thesis, we propose a physics-based method to synthesize artificial facial images in visible wavelengths from multiband NIR images. Studies on photometric properties of human skin show that melanin and hemoglobin components are dominant factors that affect the skin appearance under different light spectrum. Our physics-based synthesis method takes account of the photometric properties of human skin, introducing a skin pigment model. Specifically, a set of intensities observed at a certain surface point with varying wavelength is represented by a linear combination of both the pigment components. Our method learns the spectral basis vectors, which describe absorbance of both the pigments, from multispectral image dataset by using Independent Component Analysis (ICA). Then, our method estimates the coefficients, which are pixel-wise densities of both the pigments, from a multiband NIR image, and finally converts it to a visible light image. We demonstrate that our method works well for real facial images even though only a small dataset is available for learning basis vectors. | |||||
書誌情報 | 発行日 2013-03-25 | |||||
学位名 | ||||||
学位名 | 修士(情報理工学) | |||||
学位 | ||||||
値 | master | |||||
研究科・専攻 | ||||||
情報理工学系研究科・電子情報学専攻 | ||||||
学位授与年月日 | ||||||
学位授与年月日 | 2013-03-25 |