aiqbench analysis
This research addresses a critical gap in video analysis by proposing Label-Guided Knowledge Distillation (LGKD) specifically for 3D-CNNs, moving beyond simple adaptations of image-based methods to better handle temporal features. By integrating sample-wise and class-wise distillation components, the technique effectively transfers complex knowledge from teacher models to lightweight students while mitigating noise in action recognition tasks. The approach demonstrates improved generalization on standard benchmarks like UCF101 and HMDB51, offering a practical pathway for deploying efficient deep learning models in dynamic video environments.
Label-Guided Knowledge Distillation for 3D-CNNs in Action Recognition
arXiv:2609.13024v1 Announce Type: cross Abstract: As a key model compression technique, knowledge distillation aims to transfer knowledge from a high-capacity teacher model to a lightweight student model for enhancing the latter's performance. In this work, we reviewed the feature knowledge distillation for 3D-CNNs and observed that most feature distillation methods in video analysis are simple adaptations of those used in image analysis, often neglecting the differences of video features in the temporal dimension. To address this issue, we proposed Label-Guided Knowledge Distillation (LGKD) to guide the distillation of student model features using ground truth labels. Our method entails two components: sample-wise distillation and class-wise distillation, enabling the student model to learn feature representation of the teacher model at two levels. Sample-wise distillation utilizes label information and the teacher's probability distribution to guide the learning of features that significantly impact temporal accuracy while mitigating noise. Meanwhile, class-wise feature distillation employs a prototype network to further capture the relational knowledge among samples within the same category, enhancing the student's ability to learn higher-dimensional semantic information and improving model generalization. To demonstrate the effectiveness and superiority of our method, we conducted comprehensive experiments on two benchmark action recognition datasets, UCF101 and HMDB51, achieving competitive results.
Abstract:As a key model compression technique, knowledge distillation aims to transfer knowledge from a high-capacity teacher model to a lightweight student model for enhancing the latter's performance. In this work, we reviewed the feature knowledge distillation for 3D-CNNs and observed that most feature distillation methods in video analysis are simple adaptations of those used in image analysis, often neglecting the differences of video features in the temporal dimension. To address this issue, we proposed Label-Guided Knowledge Distillation (LGKD) to guide the distillation of student model features using ground truth labels. Our method entails two components: sample-wise distillation and class-wise distillation, enabling the student model to learn feature representation of the teacher model at two levels. Sample-wise distillation utilizes label information and the teacher's probability distribution to guide the learning of features that significantly impact temporal accuracy while mitigating noise. Meanwhile, class-wise feature distillation employs a prototype network to further capture the relational knowledge among samples within the same category, enhancing the student's ability to learn higher-dimensional semantic information and improving model generalization. To demonstrate the effectiveness and superiority of our method, we conducted comprehensive experiments on two benchmark action recognition datasets, UCF101 and HMDB51, achieving competitive results.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2609.13024 [cs.CV]
(or arXiv:2609.13024v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2609.13024 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 297 words
Related coverage
Research Note: Filtering Subversion-Relevant Information From Pretraining Data Is Feasible
LessWrong · Oct 5, 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · Oct 13, 2026
SimForcing: Distilling Simulation Motion Priors into Real-Domain Robot World Models
arXiv · Oct 7, 2026
The generative AI customization spectrum: From prompt engineering to custom models on AWS
AWS · Sep 14, 2026