Toward an Automatic System for Computer-Aided Assessment in Facial Palsy
Quantitative assessment of facial function is challenging, and subjective grading scales such as House-Brackmann, Sunnybrook, and eFACE have well-recognized limitations. Machine learning (ML) approaches to facial landmark localization carry great clinical potential as they enable high-throughput aut...
Saved in:
| Published in: | Facial plastic surgery & aesthetic medicine Vol. 22; no. 1; p. 42 |
|---|---|
| Main Authors: | , , , , , , , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
United States
01.02.2020
|
| Subjects: | |
| ISSN: | 2689-3622, 2689-3622 |
| Online Access: | Get more information |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Quantitative assessment of facial function is challenging, and subjective grading scales such as House-Brackmann, Sunnybrook, and eFACE have well-recognized limitations. Machine learning (ML) approaches to facial landmark localization carry great clinical potential as they enable high-throughput automated quantification of relevant facial metrics from photographs and videos. However, the translation from research settings to clinical application still requires important improvements.
To develop a novel ML algorithm for fast and accurate localization of facial landmarks in photographs of facial palsy patients and utilize this technology as part of an automated computer-aided diagnosis system.
Portrait photographs of 8 expressions obtained from 200 facial palsy patients and 10 healthy participants were manually annotated by localizing 68 facial landmarks in each photograph and by 3 trained clinicians using a custom graphical user interface. A novel ML model for automated facial landmark localization was trained using this disease-specific database. Algorithm accuracy was compared with manual markings and the output of a model trained using a larger database consisting only of healthy subjects.
Root mean square error normalized by the interocular distance (NRMSE) of facial landmark localization between prediction of ML algorithm and manually localized landmarks.
Publicly available algorithms for facial landmark localization provide poor localization accuracy when applied to photographs of patients compared with photographs of healthy controls (NRMSE, 8.56 ± 2.16 vs. 7.09 ± 2.34,
≪ 0.01). We found significant improvement in facial landmark localization accuracy for the facial palsy patient population when using a model trained with a relatively small number photographs (1440) of patients compared with a model trained using several thousand more images of healthy faces (NRMSE, 6.03 ± 2.43 vs. 8.56 ± 2.16,
≪ 0.01).
Retraining a computer vision facial landmark detection model with fewer than 1600 annotated images of patients significantly improved landmark detection performance in frontal view photographs of this population. The new annotated database and facial landmark localization model represent the first steps toward an automatic system for computer-aided assessment in facial palsy.
4. |
|---|---|
| AbstractList | Importance: Quantitative assessment of facial function is challenging, and subjective grading scales such as House-Brackmann, Sunnybrook, and eFACE have well-recognized limitations. Machine learning (ML) approaches to facial landmark localization carry great clinical potential as they enable high-throughput automated quantification of relevant facial metrics from photographs and videos. However, the translation from research settings to clinical application still requires important improvements. Objective: To develop a novel ML algorithm for fast and accurate localization of facial landmarks in photographs of facial palsy patients and utilize this technology as part of an automated computer-aided diagnosis system. Design, Setting, and Participants: Portrait photographs of 8 expressions obtained from 200 facial palsy patients and 10 healthy participants were manually annotated by localizing 68 facial landmarks in each photograph and by 3 trained clinicians using a custom graphical user interface. A novel ML model for automated facial landmark localization was trained using this disease-specific database. Algorithm accuracy was compared with manual markings and the output of a model trained using a larger database consisting only of healthy subjects. Main Outcomes and Measurements: Root mean square error normalized by the interocular distance (NRMSE) of facial landmark localization between prediction of ML algorithm and manually localized landmarks. Results: Publicly available algorithms for facial landmark localization provide poor localization accuracy when applied to photographs of patients compared with photographs of healthy controls (NRMSE, 8.56 ± 2.16 vs. 7.09 ± 2.34, p ≪ 0.01). We found significant improvement in facial landmark localization accuracy for the facial palsy patient population when using a model trained with a relatively small number photographs (1440) of patients compared with a model trained using several thousand more images of healthy faces (NRMSE, 6.03 ± 2.43 vs. 8.56 ± 2.16, p ≪ 0.01). Conclusions and Relevance: Retraining a computer vision facial landmark detection model with fewer than 1600 annotated images of patients significantly improved landmark detection performance in frontal view photographs of this population. The new annotated database and facial landmark localization model represent the first steps toward an automatic system for computer-aided assessment in facial palsy. Level of Evidence: 4.Importance: Quantitative assessment of facial function is challenging, and subjective grading scales such as House-Brackmann, Sunnybrook, and eFACE have well-recognized limitations. Machine learning (ML) approaches to facial landmark localization carry great clinical potential as they enable high-throughput automated quantification of relevant facial metrics from photographs and videos. However, the translation from research settings to clinical application still requires important improvements. Objective: To develop a novel ML algorithm for fast and accurate localization of facial landmarks in photographs of facial palsy patients and utilize this technology as part of an automated computer-aided diagnosis system. Design, Setting, and Participants: Portrait photographs of 8 expressions obtained from 200 facial palsy patients and 10 healthy participants were manually annotated by localizing 68 facial landmarks in each photograph and by 3 trained clinicians using a custom graphical user interface. A novel ML model for automated facial landmark localization was trained using this disease-specific database. Algorithm accuracy was compared with manual markings and the output of a model trained using a larger database consisting only of healthy subjects. Main Outcomes and Measurements: Root mean square error normalized by the interocular distance (NRMSE) of facial landmark localization between prediction of ML algorithm and manually localized landmarks. Results: Publicly available algorithms for facial landmark localization provide poor localization accuracy when applied to photographs of patients compared with photographs of healthy controls (NRMSE, 8.56 ± 2.16 vs. 7.09 ± 2.34, p ≪ 0.01). We found significant improvement in facial landmark localization accuracy for the facial palsy patient population when using a model trained with a relatively small number photographs (1440) of patients compared with a model trained using several thousand more images of healthy faces (NRMSE, 6.03 ± 2.43 vs. 8.56 ± 2.16, p ≪ 0.01). Conclusions and Relevance: Retraining a computer vision facial landmark detection model with fewer than 1600 annotated images of patients significantly improved landmark detection performance in frontal view photographs of this population. The new annotated database and facial landmark localization model represent the first steps toward an automatic system for computer-aided assessment in facial palsy. Level of Evidence: 4. Quantitative assessment of facial function is challenging, and subjective grading scales such as House-Brackmann, Sunnybrook, and eFACE have well-recognized limitations. Machine learning (ML) approaches to facial landmark localization carry great clinical potential as they enable high-throughput automated quantification of relevant facial metrics from photographs and videos. However, the translation from research settings to clinical application still requires important improvements. To develop a novel ML algorithm for fast and accurate localization of facial landmarks in photographs of facial palsy patients and utilize this technology as part of an automated computer-aided diagnosis system. Portrait photographs of 8 expressions obtained from 200 facial palsy patients and 10 healthy participants were manually annotated by localizing 68 facial landmarks in each photograph and by 3 trained clinicians using a custom graphical user interface. A novel ML model for automated facial landmark localization was trained using this disease-specific database. Algorithm accuracy was compared with manual markings and the output of a model trained using a larger database consisting only of healthy subjects. Root mean square error normalized by the interocular distance (NRMSE) of facial landmark localization between prediction of ML algorithm and manually localized landmarks. Publicly available algorithms for facial landmark localization provide poor localization accuracy when applied to photographs of patients compared with photographs of healthy controls (NRMSE, 8.56 ± 2.16 vs. 7.09 ± 2.34, ≪ 0.01). We found significant improvement in facial landmark localization accuracy for the facial palsy patient population when using a model trained with a relatively small number photographs (1440) of patients compared with a model trained using several thousand more images of healthy faces (NRMSE, 6.03 ± 2.43 vs. 8.56 ± 2.16, ≪ 0.01). Retraining a computer vision facial landmark detection model with fewer than 1600 annotated images of patients significantly improved landmark detection performance in frontal view photographs of this population. The new annotated database and facial landmark localization model represent the first steps toward an automatic system for computer-aided assessment in facial palsy. 4. |
| Author | Taati, Babak Yunusova, Yana Hadlock, Tessa A Dusseldorp, Joseph R Jowett, Nate Mohan, Suresh Tavares, Joana Fortier, Emily Guarin, Diego L van Veen, Martinus M |
| Author_xml | – sequence: 1 givenname: Diego L surname: Guarin fullname: Guarin, Diego L organization: Department of Otolaryngology/Head and Neck Surgery, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts – sequence: 2 givenname: Yana surname: Yunusova fullname: Yunusova, Yana organization: Hurvitz Brain Sciences Program, Sunnybrook Research Institute, Toronto, Canada – sequence: 3 givenname: Babak surname: Taati fullname: Taati, Babak organization: Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Canada – sequence: 4 givenname: Joseph R surname: Dusseldorp fullname: Dusseldorp, Joseph R organization: Department of Plastic and Reconstructive Surgery, Royal Australasian College of Surgeons and University of Sydney, Sydney, Australia – sequence: 5 givenname: Suresh surname: Mohan fullname: Mohan, Suresh organization: Department of Otolaryngology/Head and Neck Surgery, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts – sequence: 6 givenname: Joana surname: Tavares fullname: Tavares, Joana organization: Faculty of Health Sciences, Brasilia University, Brasilia, Brazil – sequence: 7 givenname: Martinus M surname: van Veen fullname: van Veen, Martinus M organization: Department of Otolaryngology/Head and Neck Surgery, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts – sequence: 8 givenname: Emily surname: Fortier fullname: Fortier, Emily organization: Department of Otolaryngology/Head and Neck Surgery, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts – sequence: 9 givenname: Tessa A surname: Hadlock fullname: Hadlock, Tessa A organization: Department of Otolaryngology/Head and Neck Surgery, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts – sequence: 10 givenname: Nate surname: Jowett fullname: Jowett, Nate organization: Department of Otolaryngology/Head and Neck Surgery, Massachusetts Eye and Ear Infirmary and Harvard Medical School, Boston, Massachusetts |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/32053425$$D View this record in MEDLINE/PubMed |
| BookMark | eNpNj11LwzAYRoNM3Jz7ByK59KY1ffPR5rIM54SBgvO6ZMlbqTRtbVJk_96BE7w6z8XhwHNNZl3fISG3GUszVuiHegjGp8AynYJmjKUfk7kgC1CFTrgCmP3bc7IK4fMkAQeteX5F5hyY5ALkgmz3_bcZHTUdLafYexMbS9-OIaKndT_Sde-HKeKYlI1DR8sQMASPXaRNRzfGNqalr6YNxxtyWZ-IqzOX5H3zuF9vk93L0_O63CVW6CImBw51gUrlxcEVGebCKAsaRS2URiWsZTnPmQMlFYgcHBS5ksYAdxKcYQhLcv_bHcb-a8IQK98Ei21rOuynUAGXUksmFT-pd2d1Onh01TA23ozH6u89_AD69F9V |
| CitedBy_id | crossref_primary_10_1016_j_engappai_2023_105832 crossref_primary_10_1044_2023_JSLHR_22_00321 crossref_primary_10_1109_JBHI_2020_3019242 crossref_primary_10_1055_s_0041_1726313 crossref_primary_10_1016_j_otot_2022_02_010 crossref_primary_10_1177_01945998211004169 crossref_primary_10_1097_PRS_0000000000011924 crossref_primary_10_1089_fpsam_2022_0424 crossref_primary_10_3390_healthcare10040659 crossref_primary_10_1089_fpsam_2022_0104 crossref_primary_10_1097_GOX_0000000000003638 crossref_primary_10_1155_2022_6305590 crossref_primary_10_3389_fsurg_2023_1266399 crossref_primary_10_1109_ACCESS_2023_3287389 crossref_primary_10_1089_fpsam_2024_0278 crossref_primary_10_3390_jcm11174998 crossref_primary_10_1007_s00115_020_01050_4 crossref_primary_10_1177_1357633X251342335 crossref_primary_10_1016_j_anorl_2024_07_005 crossref_primary_10_1055_s_0041_1733995 crossref_primary_10_1097_PRS_0000000000009453 crossref_primary_10_1089_fpsam_2023_0019 crossref_primary_10_1089_fpsam_2024_0381 crossref_primary_10_1016_j_otot_2021_10_014 crossref_primary_10_1155_2022_4254932 crossref_primary_10_1097_GOX_0000000000004762 crossref_primary_10_1097_PRS_0000000000010572 crossref_primary_10_1109_JBHI_2020_3045743 crossref_primary_10_1177_27325016211022805 crossref_primary_10_1002_lary_31471 crossref_primary_10_1080_02699052_2021_1890218 crossref_primary_10_3389_fnins_2025_1646485 crossref_primary_10_1089_fpsam_2022_0235 crossref_primary_10_1097_SCS_0000000000011638 crossref_primary_10_1093_ejo_cjae056 crossref_primary_10_1007_s00784_022_04724_2 crossref_primary_10_1016_j_aforl_2025_01_001 crossref_primary_10_1097_JS9_0000000000000391 crossref_primary_10_1089_fpsam_2020_0300 crossref_primary_10_3390_app12199548 crossref_primary_10_1089_fpsam_2024_0177 crossref_primary_10_1109_TNSRE_2024_3447881 crossref_primary_10_3390_jpm13071135 crossref_primary_10_1089_fpsam_2022_0282 crossref_primary_10_1016_j_bspc_2023_105942 crossref_primary_10_1016_j_bjps_2025_02_049 crossref_primary_10_1089_fpsam_2022_0364 crossref_primary_10_3390_jpm12101739 crossref_primary_10_1089_fpsam_2022_0128 crossref_primary_10_1089_fpsam_2023_0087 crossref_primary_10_1007_s00138_024_01616_1 crossref_primary_10_1159_000529685 crossref_primary_10_1016_j_irbm_2025_100882 crossref_primary_10_1097_PRS_0000000000007572 crossref_primary_10_1177_17531934251334887 crossref_primary_10_1089_fpsam_2020_0556 crossref_primary_10_3390_diagnostics12071528 crossref_primary_10_1109_JBHI_2025_3546019 crossref_primary_10_37155_2972_449X_vol1_3__79 crossref_primary_10_1109_TNSRE_2023_3244563 crossref_primary_10_1089_fpsam_2022_0136 crossref_primary_10_1109_ACCESS_2025_3550577 crossref_primary_10_1007_s10462_025_11246_2 crossref_primary_10_3390_diagnostics12051138 crossref_primary_10_1016_j_anplas_2024_11_002 crossref_primary_10_1089_fpsam_2021_0247 crossref_primary_10_1089_fpsam_2024_0204 crossref_primary_10_1109_ACCESS_2023_3330242 crossref_primary_10_3390_app11052435 crossref_primary_10_3390_app13169130 crossref_primary_10_1016_j_compbiomed_2023_107194 crossref_primary_10_3390_s21144858 |
| ContentType | Journal Article |
| DBID | CGR CUY CVF ECM EIF NPM 7X8 |
| DOI | 10.1089/fpsam.2019.29000.gua |
| DatabaseName | Medline MEDLINE MEDLINE (Ovid) MEDLINE MEDLINE PubMed MEDLINE - Academic |
| DatabaseTitle | MEDLINE Medline Complete MEDLINE with Full Text PubMed MEDLINE (Ovid) MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic MEDLINE |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | no_fulltext_linktorsrc |
| EISSN | 2689-3622 |
| ExternalDocumentID | 32053425 |
| Genre | Journal Article |
| GrantInformation_xml | – fundername: NIDCD NIH HHS grantid: R01 DC013547 |
| GroupedDBID | 0R~ ABJNI ALMA_UNASSIGNED_HOLDINGS BNQNF CGR CUY CVF ECM EIF NPM RML 7X8 SCNPE |
| ID | FETCH-LOGICAL-c498t-b32f8e6678bd81e74a6c29e4f469e64cc07370d26562472d28765aa23d52da0e2 |
| IEDL.DBID | 7X8 |
| ISICitedReferencesCount | 73 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000525077900007&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 2689-3622 |
| IngestDate | Fri Sep 05 10:47:13 EDT 2025 Thu Apr 03 06:58:47 EDT 2025 |
| IsDoiOpenAccess | false |
| IsOpenAccess | true |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 1 |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c498t-b32f8e6678bd81e74a6c29e4f469e64cc07370d26562472d28765aa23d52da0e2 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| PMID | 32053425 |
| PQID | 2355950563 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_2355950563 pubmed_primary_32053425 |
| PublicationCentury | 2000 |
| PublicationDate | 2020-02-01 |
| PublicationDateYYYYMMDD | 2020-02-01 |
| PublicationDate_xml | – month: 02 year: 2020 text: 2020-02-01 day: 01 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | Facial plastic surgery & aesthetic medicine |
| PublicationTitleAlternate | Facial Plast Surg Aesthet Med |
| PublicationYear | 2020 |
| SSID | ssj0002329937 |
| Score | 2.484007 |
| Snippet | Quantitative assessment of facial function is challenging, and subjective grading scales such as House-Brackmann, Sunnybrook, and eFACE have well-recognized... Importance: Quantitative assessment of facial function is challenging, and subjective grading scales such as House-Brackmann, Sunnybrook, and eFACE have... |
| SourceID | proquest pubmed |
| SourceType | Aggregation Database Index Database |
| StartPage | 42 |
| SubjectTerms | Adolescent Adult Aged Aged, 80 and over Anatomic Landmarks Child Diagnosis, Computer-Assisted Facial Expression Facial Paralysis - diagnosis Female Humans Machine Learning Male Middle Aged Photography |
| Title | Toward an Automatic System for Computer-Aided Assessment in Facial Palsy |
| URI | https://www.ncbi.nlm.nih.gov/pubmed/32053425 https://www.proquest.com/docview/2355950563 |
| Volume | 22 |
| WOSCitedRecordID | wos000525077900007&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV07T8MwELaAMrDwEK_ykpFY3SaO49gTihBVF6oOReoW-RXUJSkkReLfc86jTEhILN4iOafzfd-dz_ch9CBDZaXhilChA8JizYgII06UDJiLHABobhqxiWQ2E8ulnHcFt6prq-xjYhOobWl8jXxMARilh-vocf1OvGqUv13tJDR20SACKuNbupKl2NZYgC14-PX6clxIArGa9q_nhBzn60r5t-ihHFEvnTl626jfeWaDN5Oj_-70GB12TBOnrWucoB1XnKLpommTxarA6aYum3mtuJ1ajoG-4l7kgaQr6yxOt3M78arAE-Xr63gOHvt1hl4nz4unKenEFIhhUtRERzQXjgM2aStClzDFDZWO5ZAfO86MgbOeBJYCv6MsoRYyKR4rRSMbU6sCR8_RXlEW7hJhJ1zg4Owbbb2AWa5DIWKugXvkobQ6HqL73jAZOKu_gVCFKzdV9mOaIbporZut26kaWUQhHkAEufrD19fogPq8t-mevkGDHH7c3aJ981mvqo-7xgtgnc1fvgH7BLsF |
| linkProvider | ProQuest |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Toward+an+Automatic+System+for+Computer-Aided+Assessment+in+Facial+Palsy&rft.jtitle=Facial+plastic+surgery+%26+aesthetic+medicine&rft.au=Guarin%2C+Diego+L&rft.au=Yunusova%2C+Yana&rft.au=Taati%2C+Babak&rft.au=Dusseldorp%2C+Joseph+R&rft.date=2020-02-01&rft.eissn=2689-3622&rft.volume=22&rft.issue=1&rft.spage=42&rft_id=info:doi/10.1089%2Ffpsam.2019.29000.gua&rft_id=info%3Apmid%2F32053425&rft_id=info%3Apmid%2F32053425&rft.externalDocID=32053425 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=2689-3622&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=2689-3622&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=2689-3622&client=summon |