Figure 5 - uploaded by Gaurvi Goyal
Content may be subject to copyright.
Similar publications
Citations
With the global goal of carbon neutrality being emphasized, the implementation of carbon-neutral strategies has become a crucial task across various domains. As an integral part of social activities, physical education also necessitates considerations on how to reduce carbon emissions and implement carbon-neutral strategies within the teaching process. This study focuses on physical education and explores carbon-neutral strategies based on an end-to-end architecture with an attention mechanism. Firstly, we introduce an end-to-end framework that enables the integration and optimization of various aspects within the teaching process to achieve comprehensive carbon-neutral objectives. This framework serves as a unified optimization platform, facilitating the collaboration of different components involved in teaching activities and balancing the reduction of carbon emissions with teaching effectiveness. Secondly, we employ Convolutional Neural Networks (CNN) as the foundational model within the end-to-end architecture. Through training the CNN model, we automate the analysis of carbon emissions during the teaching process and provide corresponding carbon-neutral recommendations for different segments. Most importantly, we incorporate an attention mechanism to enhance the effectiveness and interpretability of the carbon-neutral strategy. The attention mechanism assists the model in automatically focusing on features or regions closely related to carbon-neutral objectives, thereby achieving more accurate and efficient carbon-neutral strategy recommendations. Finally, we conduct training and testing on the proposed model using a dataset constructed from carbon-neutral scenarios in physical education (the country where physical education occurred and digital energy have been scrutinized). The results demonstrate that the improved model surpasses a 90% threshold in mainstream evaluation metrics such as Action Recognition Accuracy (ARA), Action Recognition Recall (ARR), and Action Optimization Rate (AOR). The enhanced model exhibits notable improvements in inference speed and accuracy.
The widespread use of visual surveillance in public areas puts individual privacy at stake while also increasing resource usage (energy, bandwidth, and computation). Neuromorphic vision sensors (or event cameras) are considered viable solutions for privacy issues; since event cameras only capture scene dynamics, they do not capture detailed RGB images of individuals. However, recent deep learning architectures have enabled the reconstruction of high-fidelity images from event sensor data that could reveal individual identity information. As a result, it reintroduces privacy risks for event-based vision applications. In this work, we focus on protecting the identity of individuals from such image reconstruction attacks by anonymizing event data. To achieve this, we present an end-to-end network architecture jointly optimized for the twofold objective of preserving privacy and performing a downstream computer vision task. The proposed network learns to scramble events, thereby degrading the quality of images that potential intruders could reconstruct. We demonstrate the application of our framework in two challenging computer vision tasks: person re-identification (ReId) and human pose estimation (HPE). To this end, we also present and evaluate the first event-based person ReId dataset, Event-ReId. We validate the privacy-preserving efficacy of our approach against possible privacy attacks through extensive experiments: for person ReId, we utilize the real event-based Event-ReId dataset and synthetic event data simulated from the SoftBio dataset; for HPE, we use a publicly available event-based dataset DHP19. The results of both tasks show that anonymizing event data effectively protects private information with minimal impact on the subsequent task performance.
