{"id":14657,"date":"2023-01-07T01:16:53","date_gmt":"2023-01-07T01:16:53","guid":{"rendered":"https:\/\/spa-g.com\/?p=14657"},"modified":"2024-05-07T15:03:08","modified_gmt":"2024-05-07T08:03:08","slug":"seminar-bieu-dien-anh-bang-cac-vector-dac-trung-chieu-cao-toi-uu-do-thi-emr-va-tinh-toan-xep-hang-anh-trong-cbir","status":"publish","type":"post","link":"https:\/\/spa-g.com\/seminar-bieu-dien-anh-bang-cac-vector-dac-trung-chieu-cao-toi-uu-do-thi-emr-va-tinh-toan-xep-hang-anh-trong-cbir\/","title":{"rendered":"Seminar Bi\u1ec3u di\u1ec5n \u1ea3nh b\u1eb1ng c\u00e1c vector \u0111\u1eb7c tr\u01b0ng chi\u1ec1u cao, t\u1ed1i \u01b0u \u0111\u1ed3 th\u1ecb EMR v\u00e0 t\u00ednh to\u00e1n x\u1ebfp h\u1ea1ng \u1ea3nh trong CBIR"},"content":{"rendered":"
Ng\u00e0y 05\/01 v\u1eeba qua, Khoa C\u00f4ng ngh\u1ec7 Th\u00f4ng tin & Truy\u1ec1n th\u00f4ng t\u1ed5 ch\u1ee9c seminar Bi\u1ec3u di\u1ec5n \u1ea3nh b\u1eb1ng c\u00e1c vector \u0111\u1eb7c tr\u01b0ng chi\u1ec1u cao, t\u1ed1i \u01b0u \u0111\u1ed3 th\u1ecb EMR (X\u1ebfp h\u1ea1ng \u0111a t\u1ea1p hi\u1ec7u qu\u1ea3) v\u00e0 t\u00ednh to\u00e1n x\u1ebfp h\u1ea1ng \u1ea3nh trong CBIR (Truy v\u1ea5n \u1ea3nh theo n\u1ed9i dung) v\u1edbi s\u1ef1 tham gia c\u1ee7a c\u00e1c th\u1ea7y c\u00f4 h\u1ed9i \u0111\u1ed3ng chuy\u00ean m\u00f4n khoa CNTT&TT v\u00e0 c\u00e1c nh\u00e0 khoa h\u1ecdc quan t\u00e2m \u0111\u1ebfn ch\u1ee7 \u0111\u1ec1 b\u00e1o c\u00e1o.<\/i><\/b><\/p>\n
Tham d\u1ef1 s\u1ef1 ki\u1ec7n c\u00f3 s\u1ef1 c\u00f3 m\u1eb7t c\u1ee7a TS Nguy\u1ec5n Ng\u1ecdc B\u00ecnh – Hi\u1ec7u tr\u01b0\u1edfng Tr\u01b0\u1eddng \u0110\u1ea1i h\u1ecdc CMC, TS. Nguy\u1ec5n Kim C\u01b0\u01a1ng \u2013 Tr\u01b0\u1edfng ban \u0110\u1ea1i h\u1ecdc S\u1ed1,\u00a0PGS.TS B\u00f9i Thu L\u00e2m \u2013 Ch\u1ee7 t\u1ecbch H\u1ed9i \u0111\u1ed3ng Chuy\u00ean m\u00f4n Khoa; PGS.TS \u0110\u1ed7 V\u0103n Th\u00e0nh \u2013 Gi\u00e1m \u0111\u1ed1c Ch\u01b0\u01a1ng tr\u00ecnh \u0110\u00e0o t\u1ea1o C\u00f4ng ngh\u1ec7 Th\u00f4ng tin, Ph\u00f3 Ch\u1ee7 t\u1ecbch H\u1ed9i \u0111\u1ed3ng Chuy\u00ean m\u00f4n Khoa v\u1edbi s\u1ef1 tr\u00ecnh b\u00e0y c\u1ee7a TS Ng\u00f4 Ho\u00e0ng Huy – Tr\u01b0\u1eddng \u0110\u1ea1i h\u1ecdc CMC v\u00e0 ThS\u00a0<\/span>Ho\u00e0ng V\u0103n Qu\u00fd – Tr\u01b0\u1eddng \u0110\u1ea1i h\u1ecdc H\u1ed3ng \u0110\u1ee9c.<\/p>\n <\/p>\n N\u1ed9i dung seminar tr\u00ecnh b\u00e0y 2 k\u1ebft qu\u1ea3 nghi\u00ean c\u1ee9u m\u1edbi v\u1ec1 CBIR d\u1ef1a tr\u00ean thu\u1eadt to\u00e1n EMR:\u00a0<\/span><\/p>\n V\u1ea5n \u0111\u1ec1 1, x\u00e1c \u0111\u1ecbnh c\u00e1c anchor points c\u1ee7a EMR d\u1ef1a tr\u00ean thu\u1eadt to\u00e1n ph\u00e2n c\u1ee5m m\u1edbi ki\u1ec3u C-means v\u1edbi \u0111i\u1ec1u ki\u1ec7n d\u1eef li\u1ec7u l\u1edbn (s\u1ed1 vector, s\u1ed1 c\u1ee5m v\u00e0 s\u1ed1 chi\u1ec1u c\u1ee7a c\u00e1c vector \u0111\u1ec1u r\u1ea5t l\u1edbn). Thu\u1eadt to\u00e1n (t\u1ea1m g\u1ecdi l\u00e0 LDM-FCM) \u0111\u01b0\u1ee3c x\u00e2y d\u1ef1ng v\u1edbi t\u00ednh to\u00e1n song song tr\u00ean GPU, s\u1eed d\u1ee5ng k\u1ef9 thu\u1eadt t\u00ecm ki\u1ebfm vector ANN (approximate nearest neighbor) \u0111\u1ec3 t\u1ed1i \u01b0u x\u1eed l\u00fd ph\u00e2n c\u1ee5m C-means tr\u00ean t\u1eadp d\u1eef li\u1ec7u vector \u0111\u1eb7c tr\u01b0ng ki\u1ec3u local data manifold (c\u00e1c vector \u0111\u01b0\u1ee3c k\u1ebft h\u1ee3p t\u1eeb c\u00e1c \u0111\u1eb7c tr\u01b0ng m\u1ee9c th\u1ea5p v\u00e0 c\u00e1c \u0111\u1eb7c tr\u01b0ng tr\u00edch xu\u1ea5t t\u1eeb c\u00e1c m\u1ea1ng h\u1ecdc s\u00e2u \u0111\u00e3 hu\u1ea5n luy\u1ec7n\u00a0 v\u00e0 c\u00f3 k\u1ebft h\u1ee3p k\u1ef9 thu\u1eadt fine-turning).<\/span><\/p>\n V\u1ea5n \u0111\u1ec1 2, t\u1ed1i \u01b0u qu\u00e1 tr\u00ecnh x\u00e2y d\u1ef1ng \u0111\u1ed3 th\u1ecb EMR v\u00e0 t\u00ednh to\u00e1n x\u1ebfp h\u1ea1ng dataset \u1ea3nh theo \u1ea3nh truy v\u1ea5n khi \u1ea3nh \u0111\u01b0\u1ee3c bi\u1ec3u di\u1ec5n b\u1eb1ng c\u00e1c vector chi\u1ec1u cao v\u00e0 s\u1ed1 anchor points r\u1ea5t l\u1edbn. Thu\u1eadt to\u00e1n (t\u1ea1m g\u1ecdi l\u00e0 HD-EMR) \u0111\u00e3 s\u1eed d\u1ee5ng LDM-FCM \u0111\u1ec3 x\u00e1c \u0111\u1ecbnh anchor points c\u1ee7a EMR,\u0111\u1ed3ng th\u1eddi s\u1eed d\u1ee5ng\u00a0 k\u1ef9 thu\u1eadt t\u00ecm ki\u1ebfm vector ANN \u0111\u1ec3 t\u00ednh c\u00e1c ma tr\u1eadn bi\u1ec3u di\u1ec5n \u0111\u1ed3 th\u1ecb quan h\u1ec7 gi\u1eefa c\u00e1c vector \u1ea3nh c\u1ee7a EMR.<\/span><\/p>\n <\/p>\n Cu\u1ed1i c\u00f9ng m\u1ed9t th\u1ee7 t\u1ee5c l\u1eb7p t\u00ednh to\u00e1n tr\u00ean c\u00e1c ma tr\u1eadn th\u01b0a, kh\u00f4ng s\u1eed d\u1ee5ng c\u00e1c ma tr\u1eadn ngh\u1ecbch \u0111\u1ea3o c\u1ee1 l\u1edbn, t\u1ed1i \u01b0u v\u1ec1 t\u00e0i nguy\u00ean (b\u1ed9 nh\u1edb v\u00e0 th\u1eddi gian) \u0111\u01b0\u1ee3c thi\u1ebft l\u1eadp \u0111\u1ec3 x\u1ebfp h\u1ea1ng dataset \u1ea3nh theo \u1ea3nh truy v\u1ea5n.<\/span><\/p>\n Hi\u1ec7n nay ngo\u00e0i c\u00e1c b\u1ed9 m\u00f4 t\u1ea3 \u1ea3nh theo low-level feature (\u0111\u1eb7c tr\u01b0ng m\u1ee9c th\u1ea5p, LF) truy\u1ec1n th\u1ed1ng, c\u00e1c b\u1ed9 m\u00f4 t\u1ea3 \u1ea3nh b\u1eb1ng c\u00e1c Deep feature embedding (DFE) \u0111\u00e3 \u0111\u01b0\u1ee3c \u1ee9ng d\u1ee5ng r\u1ed9ng r\u00e3i v\u00e0 c\u00f3 hi\u1ec7u su\u1ea5t r\u1ea5t cao trong truy v\u1ea5n \u1ea3nh theo n\u1ed9i dung (CBIR).<\/span><\/p>\n <\/p>\n M\u1ed9t dataset, qua c\u00e1c b\u1ed9 m\u00f4 t\u1ea3 \u1ea3nh DFE s\u1ebd \u0111\u01b0\u1ee3c xem nh\u01b0 l\u00e0 m\u1ed9t local data manifold v\u1edbi vector bi\u1ec3u di\u1ec5n l\u00e0 chi\u1ec1u r\u1ea5t cao, v\u00e0 do \u0111\u1eb7c t\u00ednh \u0111a t\u1ea1p c\u00e1c \u0111\u1ed9 \u0111o kho\u1ea3ng c\u00e1ch global ki\u1ec3u Euclidean, cosin s\u1ebd \u0111\u01b0\u1ee3c thay th\u1ebf b\u1eb1ng c\u00e1c \u0111\u1ed9 \u0111o local \u0111\u1ec3 \u0111o \u0111\u1ed9 t\u01b0\u01a1ng t\u1ef1 \u1ea3nh.<\/span><\/p>\n EMR (x\u1ebfp h\u1ea1ng \u0111a t\u1ea1p hi\u1ec7u qu\u1ea3)l\u00e0 m\u1ed9t thu\u1eadt to\u00e1n x\u1ebfp h\u1ea1ng theo ti\u1ebfp c\u1eadn data manifold \u0111\u01b0\u1ee3c \u1ee9ng d\u1ee5ng r\u1ed9ng r\u00e3i trong CBIR, b\u1eb1ng c\u00e1ch bi\u1ec3u di\u1ec5n quan h\u1ec7 t\u01b0\u01a1ng t\u1ef1 gi\u1eefa c\u00e1c vectors th\u00f4ng qua m\u1ed9t anchor point chung trong l\u00e2n c\u1eadn, t\u00ednh “cong” c\u1ee7a manifold \u0111\u00e3 \u0111\u01b0\u1ee3c bi\u1ec3u di\u1ec5n t\u1ed1t v\u1edbi EMR.\u00a0<\/span><\/p>\n <\/p>\n Tuy v\u1eady EMR \u0111\u01b0\u1ee3c \u1ee9ng d\u1ee5ng ch\u1ee7 y\u1ebfu cho \u1ea3nh v\u1edbi c\u00e1c bi\u1ec3u di\u1ec5n LF, v\u00e0 gi\u1ea3 \u0111\u1ecbnh r\u1eb1ng s\u1ed1 anchor vectors kh\u00f4ng qu\u00e1 l\u1edbn. Do \u0111\u00f3 \u0111\u1ec3 c\u00f3 th\u1ec3 \u1ee9ng d\u1ee5ng \u0111\u01b0\u1ee3c EMR cho x\u1ebfp h\u1ea1ng c\u00e1c local data manifold bi\u1ec3u di\u1ec5n dataset \u1ea3nh nh\u00fang trong kh\u00f4ng gian Euclidean chi\u1ec1u r\u1ea5t cao,\u00a0 \u0111\u1ed3 th\u1ecb EMR v\u00e0 x\u1ebfp h\u1ea1ng hi\u1ec7u qu\u1ea3 c\u1ea7n \u0111\u01b0\u1ee3c c\u1ea3i ti\u1ebfn b\u01b0\u1edbc v\u1ec1 tr\u00edch ch\u1ecdn anchor points, t\u1ed1i \u01b0u b\u1ed9 nh\u1edb v\u00e0 th\u1eddi gian t\u00ednh to\u00e1n.<\/span><\/p>\n <\/p>\n","protected":false},"excerpt":{"rendered":" Ng\u00e0y 05\/01 v\u1eeba qua, Khoa C\u00f4ng ngh\u1ec7 Th\u00f4ng tin & Truy\u1ec1n th\u00f4ng t\u1ed5 ch\u1ee9c seminar Bi\u1ec3u di\u1ec5n \u1ea3nh b\u1eb1ng c\u00e1c vector \u0111\u1eb7c tr\u01b0ng chi\u1ec1u cao, t\u1ed1i \u01b0u \u0111\u1ed3 th\u1ecb EMR (X\u1ebfp h\u1ea1ng \u0111a t\u1ea1p hi\u1ec7u qu\u1ea3) v\u00e0 t\u00ednh to\u00e1n x\u1ebfp h\u1ea1ng \u1ea3nh trong CBIR (Truy v\u1ea5n \u1ea3nh theo n\u1ed9i dung) v\u1edbi s\u1ef1 tham gia c\u1ee7a c\u00e1c th\u1ea7y c\u00f4 h\u1ed9i \u0111\u1ed3ng chuy\u00ean m\u00f4n khoa CNTT&TT v\u00e0 c\u00e1c nh\u00e0 khoa h\u1ecdc quan t\u00e2m \u0111\u1ebfn ch\u1ee7 \u0111\u1ec1 b\u00e1o c\u00e1o. Tham d\u1ef1 s\u1ef1 ki\u1ec7n c\u00f3 s\u1ef1 c\u00f3 m\u1eb7t c\u1ee7a TS Nguy\u1ec5n Ng\u1ecdc B\u00ecnh – Hi\u1ec7u tr\u01b0\u1edfng Tr\u01b0\u1eddng \u0110\u1ea1i h\u1ecdc CMC, TS. Nguy\u1ec5n Kim C\u01b0\u01a1ng \u2013 Tr\u01b0\u1edfng ban \u0110\u1ea1i h\u1ecdc S\u1ed1,\u00a0PGS.TS B\u00f9i Thu L\u00e2m \u2013 Ch\u1ee7 t\u1ecbch H\u1ed9i \u0111\u1ed3ng Chuy\u00ean m\u00f4n Khoa; PGS.TS \u0110\u1ed7 V\u0103n Th\u00e0nh \u2013 Gi\u00e1m \u0111\u1ed1c Ch\u01b0\u01a1ng tr\u00ecnh \u0110\u00e0o t\u1ea1o C\u00f4ng ngh\u1ec7 Th\u00f4ng tin, Ph\u00f3 Ch\u1ee7 t\u1ecbch H\u1ed9i \u0111\u1ed3ng Chuy\u00ean m\u00f4n Khoa v\u1edbi s\u1ef1 tr\u00ecnh b\u00e0y c\u1ee7a TS Ng\u00f4 Ho\u00e0ng Huy – Tr\u01b0\u1eddng \u0110\u1ea1i h\u1ecdc CMC v\u00e0 ThS\u00a0Ho\u00e0ng V\u0103n Qu\u00fd – Tr\u01b0\u1eddng \u0110\u1ea1i h\u1ecdc H\u1ed3ng \u0110\u1ee9c. N\u1ed9i dung seminar tr\u00ecnh b\u00e0y 2 k\u1ebft qu\u1ea3 nghi\u00ean c\u1ee9u m\u1edbi v\u1ec1 CBIR d\u1ef1a tr\u00ean thu\u1eadt to\u00e1n EMR:\u00a0 V\u1ea5n \u0111\u1ec1 1, x\u00e1c \u0111\u1ecbnh c\u00e1c anchor points c\u1ee7a EMR d\u1ef1a tr\u00ean thu\u1eadt to\u00e1n ph\u00e2n c\u1ee5m m\u1edbi ki\u1ec3u C-means v\u1edbi \u0111i\u1ec1u ki\u1ec7n d\u1eef li\u1ec7u l\u1edbn (s\u1ed1 vector, s\u1ed1 c\u1ee5m v\u00e0 s\u1ed1 chi\u1ec1u c\u1ee7a c\u00e1c vector \u0111\u1ec1u r\u1ea5t l\u1edbn). Thu\u1eadt to\u00e1n (t\u1ea1m g\u1ecdi l\u00e0 LDM-FCM) \u0111\u01b0\u1ee3c x\u00e2y d\u1ef1ng v\u1edbi t\u00ednh to\u00e1n song song tr\u00ean GPU, s\u1eed d\u1ee5ng k\u1ef9 thu\u1eadt t\u00ecm ki\u1ebfm vector ANN (approximate nearest neighbor) \u0111\u1ec3 t\u1ed1i \u01b0u x\u1eed l\u00fd ph\u00e2n c\u1ee5m C-means tr\u00ean t\u1eadp d\u1eef li\u1ec7u vector \u0111\u1eb7c tr\u01b0ng ki\u1ec3u local data manifold (c\u00e1c vector \u0111\u01b0\u1ee3c k\u1ebft h\u1ee3p t\u1eeb c\u00e1c \u0111\u1eb7c tr\u01b0ng m\u1ee9c th\u1ea5p v\u00e0 c\u00e1c \u0111\u1eb7c tr\u01b0ng tr\u00edch xu\u1ea5t t\u1eeb c\u00e1c m\u1ea1ng h\u1ecdc s\u00e2u \u0111\u00e3 hu\u1ea5n luy\u1ec7n\u00a0 v\u00e0 c\u00f3 k\u1ebft h\u1ee3p k\u1ef9 thu\u1eadt fine-turning). V\u1ea5n \u0111\u1ec1 2, t\u1ed1i \u01b0u qu\u00e1 tr\u00ecnh x\u00e2y d\u1ef1ng \u0111\u1ed3 th\u1ecb EMR v\u00e0 t\u00ednh to\u00e1n x\u1ebfp h\u1ea1ng dataset \u1ea3nh theo \u1ea3nh truy v\u1ea5n khi \u1ea3nh \u0111\u01b0\u1ee3c bi\u1ec3u di\u1ec5n b\u1eb1ng c\u00e1c vector chi\u1ec1u cao v\u00e0 s\u1ed1 anchor points r\u1ea5t l\u1edbn. Thu\u1eadt to\u00e1n (t\u1ea1m g\u1ecdi l\u00e0 HD-EMR) \u0111\u00e3 s\u1eed d\u1ee5ng LDM-FCM \u0111\u1ec3 x\u00e1c \u0111\u1ecbnh anchor points c\u1ee7a EMR,\u0111\u1ed3ng th\u1eddi s\u1eed d\u1ee5ng\u00a0 k\u1ef9 thu\u1eadt t\u00ecm ki\u1ebfm vector ANN \u0111\u1ec3 t\u00ednh c\u00e1c ma tr\u1eadn bi\u1ec3u di\u1ec5n \u0111\u1ed3 th\u1ecb quan h\u1ec7 gi\u1eefa c\u00e1c vector \u1ea3nh c\u1ee7a EMR. Cu\u1ed1i c\u00f9ng m\u1ed9t th\u1ee7 t\u1ee5c l\u1eb7p t\u00ednh to\u00e1n tr\u00ean c\u00e1c ma tr\u1eadn th\u01b0a, kh\u00f4ng s\u1eed d\u1ee5ng c\u00e1c ma tr\u1eadn ngh\u1ecbch \u0111\u1ea3o c\u1ee1 l\u1edbn, t\u1ed1i \u01b0u v\u1ec1 t\u00e0i nguy\u00ean (b\u1ed9 nh\u1edb v\u00e0 th\u1eddi gian) \u0111\u01b0\u1ee3c thi\u1ebft l\u1eadp \u0111\u1ec3 x\u1ebfp h\u1ea1ng dataset \u1ea3nh theo \u1ea3nh truy v\u1ea5n. Hi\u1ec7n nay ngo\u00e0i c\u00e1c b\u1ed9 m\u00f4 t\u1ea3 \u1ea3nh theo low-level feature (\u0111\u1eb7c tr\u01b0ng m\u1ee9c th\u1ea5p, LF) truy\u1ec1n th\u1ed1ng, c\u00e1c b\u1ed9 m\u00f4 t\u1ea3 \u1ea3nh b\u1eb1ng c\u00e1c Deep feature embedding (DFE) \u0111\u00e3 \u0111\u01b0\u1ee3c \u1ee9ng d\u1ee5ng r\u1ed9ng r\u00e3i v\u00e0 c\u00f3 hi\u1ec7u su\u1ea5t r\u1ea5t cao trong truy v\u1ea5n \u1ea3nh theo n\u1ed9i dung (CBIR). M\u1ed9t dataset, qua c\u00e1c b\u1ed9 m\u00f4 t\u1ea3 \u1ea3nh DFE s\u1ebd \u0111\u01b0\u1ee3c xem nh\u01b0 l\u00e0 m\u1ed9t local data manifold v\u1edbi vector bi\u1ec3u di\u1ec5n l\u00e0 chi\u1ec1u r\u1ea5t cao, v\u00e0 do \u0111\u1eb7c t\u00ednh \u0111a t\u1ea1p c\u00e1c \u0111\u1ed9 \u0111o kho\u1ea3ng c\u00e1ch global ki\u1ec3u Euclidean, cosin s\u1ebd \u0111\u01b0\u1ee3c thay th\u1ebf b\u1eb1ng c\u00e1c \u0111\u1ed9 \u0111o local \u0111\u1ec3 \u0111o \u0111\u1ed9 t\u01b0\u01a1ng t\u1ef1 \u1ea3nh. EMR (x\u1ebfp h\u1ea1ng \u0111a t\u1ea1p hi\u1ec7u qu\u1ea3)l\u00e0 m\u1ed9t thu\u1eadt to\u00e1n x\u1ebfp h\u1ea1ng theo ti\u1ebfp c\u1eadn data manifold \u0111\u01b0\u1ee3c \u1ee9ng d\u1ee5ng r\u1ed9ng r\u00e3i trong CBIR, b\u1eb1ng c\u00e1ch bi\u1ec3u di\u1ec5n quan h\u1ec7 t\u01b0\u01a1ng t\u1ef1 gi\u1eefa c\u00e1c vectors th\u00f4ng qua m\u1ed9t anchor point chung trong l\u00e2n c\u1eadn, t\u00ednh “cong” c\u1ee7a manifold \u0111\u00e3 \u0111\u01b0\u1ee3c bi\u1ec3u di\u1ec5n t\u1ed1t v\u1edbi EMR.\u00a0 Tuy v\u1eady EMR \u0111\u01b0\u1ee3c \u1ee9ng d\u1ee5ng ch\u1ee7 y\u1ebfu cho \u1ea3nh v\u1edbi c\u00e1c bi\u1ec3u di\u1ec5n LF, v\u00e0 gi\u1ea3 \u0111\u1ecbnh r\u1eb1ng s\u1ed1 anchor vectors kh\u00f4ng qu\u00e1 l\u1edbn. Do \u0111\u00f3 \u0111\u1ec3 c\u00f3 th\u1ec3 \u1ee9ng d\u1ee5ng \u0111\u01b0\u1ee3c EMR cho x\u1ebfp h\u1ea1ng c\u00e1c local data manifold bi\u1ec3u di\u1ec5n dataset \u1ea3nh nh\u00fang trong kh\u00f4ng gian Euclidean chi\u1ec1u r\u1ea5t cao,\u00a0 \u0111\u1ed3 th\u1ecb EMR v\u00e0 x\u1ebfp h\u1ea1ng hi\u1ec7u qu\u1ea3 c\u1ea7n \u0111\u01b0\u1ee3c c\u1ea3i ti\u1ebfn b\u01b0\u1edbc v\u1ec1 tr\u00edch ch\u1ecdn anchor points, t\u1ed1i \u01b0u b\u1ed9 nh\u1edb v\u00e0 th\u1eddi gian t\u00ednh to\u00e1n. <\/p>\n","protected":false},"author":6,"featured_media":17144,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[640,128,124],"tags":[],"yoast_head":"\n