{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/dynamic-attention-controlled-cascaded-shape","title":"Dynamic Attention-controlled Cascaded Shape Regression Exploiting Training Data Augmentation and Fuzzy-set Sample Weighting","arxiv_id":"1611.05396","date":"2016-11-16","proceeding":"CVPR 2017 7","authors":["Zhen-Hua Feng","Josef Kittler","William Christmas","Patrik Huber","Xiao-Jun Wu"],"abstract":"We present a new Cascaded Shape Regression (CSR) architecture, namely Dynamic\nAttention-Controlled CSR (DAC-CSR), for robust facial landmark detection on\nunconstrained faces. Our DAC-CSR divides facial landmark detection into three\ncascaded sub-tasks: face bounding box refinement, general CSR and\nattention-controlled CSR. The first two stages refine initial face bounding\nboxes and output intermediate facial landmarks. Then, an online dynamic model\nselection method is used to choose appropriate domain-specific CSRs for further\nlandmark refinement. The key innovation of our DAC-CSR is the fault-tolerant\nmechanism, using fuzzy set sample weighting for attention-controlled\ndomain-specific model training. Moreover, we advocate data augmentation with a\nsimple but effective 2D profile face generator, and context-aware feature\nextraction for better facial feature representation. Experimental results\nobtained on challenging datasets demonstrate the merits of our DAC-CSR over the\nstate-of-the-art.","url_abs":"http://arxiv.org/abs/1611.05396v2","url_pdf":"http://arxiv.org/pdf/1611.05396v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"data-augmentation","task_name":"Data Augmentation"},{"task_slug":"face-alignment","task_name":"Face Alignment"},{"task_slug":"facial-landmark-detection","task_name":"Facial Landmark Detection"},{"task_slug":"model-selection","task_name":"Model Selection"},{"task_slug":"regression-1","task_name":"regression"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/face-alignment-on-aflw-19","task":"Face Alignment","dataset":"AFLW-19","model":"DAC-CSR","rank_in_archive_order":18,"of":23,"metrics":{"NME_diag (%, Frontal)":"1.81","NME_diag (%, Full)":"2.21"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1611.05396","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}