Matches in SemOpenAlex for { <https://semopenalex.org/work/W3212280470> ?p ?o ?g. }
- W3212280470 abstract "Facial landmark detection is a vital step for numerous facial image analysis applications. Although some deep learning-based methods have achieved good performances in this task, they are often not suitable for running on mobile devices. Such methods rely on networks with many parameters, which makes the training and inference time-consuming. Training lightweight neural networks such as MobileNets are often challenging, and the models might have low accuracy. Inspired by knowledge distillation (KD), this paper presents a novel loss function to train a lightweight Student network (e.g., MobileNetV2) for facial landmark detection. We use two Teacher networks, a Tolerant-Teacher and a Tough-Teacher in conjunction with the Student network. The Tolerant-Teacher is trained using Soft-landmarks created by active shape models, while the Tough-Teacher is trained using the ground truth (aka Hard-landmarks) landmark points. To utilize the facial landmark points predicted by the Teacher networks, we define an Assistive Loss (ALoss) for each Teacher network. Moreover, we define a loss function called KD-Loss that utilizes the facial landmark points predicted by the two pre-trained Teacher networks (EfficientNet-b3) to guide the lightweight Student network towards predicting the Hard-landmarks. Our experimental results on three challenging facial datasets show that the proposed architecture will result in a better-trained Student network that can extract facial landmark points with high accuracy." @default.
- W3212280470 created "2021-11-22" @default.
- W3212280470 creator A5008504935 @default.
- W3212280470 creator A5041948053 @default.
- W3212280470 date "2021-11-13" @default.
- W3212280470 modified "2023-09-27" @default.
- W3212280470 title "Facial Landmark Points Detection Using Knowledge Distillation-Based Neural Networks." @default.
- W3212280470 cites W1522301498 @default.
- W3212280470 cites W1796263212 @default.
- W3212280470 cites W1821462560 @default.
- W3212280470 cites W1916406603 @default.
- W3212280470 cites W1960706641 @default.
- W3212280470 cites W1977821862 @default.
- W3212280470 cites W1989702938 @default.
- W3212280470 cites W2019177772 @default.
- W3212280470 cites W2047508432 @default.
- W3212280470 cites W2087681821 @default.
- W3212280470 cites W2101866605 @default.
- W3212280470 cites W2109434346 @default.
- W3212280470 cites W2111372597 @default.
- W3212280470 cites W2133444763 @default.
- W3212280470 cites W2151802865 @default.
- W3212280470 cites W2175678410 @default.
- W3212280470 cites W2554268477 @default.
- W3212280470 cites W2736671157 @default.
- W3212280470 cites W2736728583 @default.
- W3212280470 cites W2746753454 @default.
- W3212280470 cites W2770121394 @default.
- W3212280470 cites W2789049214 @default.
- W3212280470 cites W2798730128 @default.
- W3212280470 cites W2802717621 @default.
- W3212280470 cites W2894332907 @default.
- W3212280470 cites W2915352401 @default.
- W3212280470 cites W2944892540 @default.
- W3212280470 cites W2962890819 @default.
- W3212280470 cites W2963544187 @default.
- W3212280470 cites W2963566548 @default.
- W3212280470 cites W2963716477 @default.
- W3212280470 cites W2963789946 @default.
- W3212280470 cites W2963993755 @default.
- W3212280470 cites W2964014798 @default.
- W3212280470 cites W2980791196 @default.
- W3212280470 cites W2982772166 @default.
- W3212280470 cites W2995802584 @default.
- W3212280470 cites W3010622397 @default.
- W3212280470 cites W3041340374 @default.
- W3212280470 cites W3109891286 @default.
- W3212280470 cites W3160676567 @default.
- W3212280470 cites W3183413007 @default.
- W3212280470 hasPublicationYear "2021" @default.
- W3212280470 type Work @default.
- W3212280470 sameAs 3212280470 @default.
- W3212280470 citedByCount "0" @default.
- W3212280470 crossrefType "posted-content" @default.
- W3212280470 hasAuthorship W3212280470A5008504935 @default.
- W3212280470 hasAuthorship W3212280470A5041948053 @default.
- W3212280470 hasConcept C108583219 @default.
- W3212280470 hasConcept C119857082 @default.
- W3212280470 hasConcept C121158502 @default.
- W3212280470 hasConcept C14036430 @default.
- W3212280470 hasConcept C144024400 @default.
- W3212280470 hasConcept C153180895 @default.
- W3212280470 hasConcept C154945302 @default.
- W3212280470 hasConcept C161191863 @default.
- W3212280470 hasConcept C2776214188 @default.
- W3212280470 hasConcept C2779304628 @default.
- W3212280470 hasConcept C2780297707 @default.
- W3212280470 hasConcept C31972630 @default.
- W3212280470 hasConcept C36289849 @default.
- W3212280470 hasConcept C41008148 @default.
- W3212280470 hasConcept C50644808 @default.
- W3212280470 hasConcept C78458016 @default.
- W3212280470 hasConcept C86803240 @default.
- W3212280470 hasConceptScore W3212280470C108583219 @default.
- W3212280470 hasConceptScore W3212280470C119857082 @default.
- W3212280470 hasConceptScore W3212280470C121158502 @default.
- W3212280470 hasConceptScore W3212280470C14036430 @default.
- W3212280470 hasConceptScore W3212280470C144024400 @default.
- W3212280470 hasConceptScore W3212280470C153180895 @default.
- W3212280470 hasConceptScore W3212280470C154945302 @default.
- W3212280470 hasConceptScore W3212280470C161191863 @default.
- W3212280470 hasConceptScore W3212280470C2776214188 @default.
- W3212280470 hasConceptScore W3212280470C2779304628 @default.
- W3212280470 hasConceptScore W3212280470C2780297707 @default.
- W3212280470 hasConceptScore W3212280470C31972630 @default.
- W3212280470 hasConceptScore W3212280470C36289849 @default.
- W3212280470 hasConceptScore W3212280470C41008148 @default.
- W3212280470 hasConceptScore W3212280470C50644808 @default.
- W3212280470 hasConceptScore W3212280470C78458016 @default.
- W3212280470 hasConceptScore W3212280470C86803240 @default.
- W3212280470 hasLocation W32122804701 @default.
- W3212280470 hasOpenAccess W3212280470 @default.
- W3212280470 hasPrimaryLocation W32122804701 @default.
- W3212280470 hasRelatedWork W2784045851 @default.
- W3212280470 hasRelatedWork W2963882342 @default.
- W3212280470 hasRelatedWork W2963915229 @default.
- W3212280470 hasRelatedWork W2964135029 @default.
- W3212280470 hasRelatedWork W2982276691 @default.
- W3212280470 hasRelatedWork W3020895012 @default.
- W3212280470 hasRelatedWork W3035472678 @default.