Výsledky vyhledávání

Report

TaQ-DiT: Time-aware Quantization for Diffusion Transformers

Autor: Liu, Xinyan, Shi, Huihong, Xu, Yang, Wang, Zhongfeng

Transformer-based diffusion models, dubbed Diffusion Transformers (DiTs), have achieved state-of-the-art performance in image and video generation tasks. However, their large model size and slow inference speed limit their practical applications, cal

Externí odkaz: http://arxiv.org/abs/2411.14172

Zobrazit plný text záznamu

Report

Weakly Supervised Video Individual CountingWeakly Supervised Video Individual Counting

Autor: Liu, Xinyan, Li, Guorong, Qi, Yuankai, Yan, Ziheng, Han, Zhenjun, Hengel, Anton van den, Yang, Ming-Hsuan, Huang, Qingming

Video Individual Counting (VIC) aims to predict the number of unique individuals in a single video. % Existing methods learn representations based on trajectory labels for individuals, which are annotation-expensive. % To provide a more realistic ref

Externí odkaz: http://arxiv.org/abs/2312.05923

Zobrazit plný text záznamu

Report

Progressive Multi-resolution Loss for Crowd Counting

Autor: Yan, Ziheng, Qi, Yuankai, Li, Guorong, Liu, Xinyan, Zhang, Weigang, Huang, Qingming, Yang, Ming-Hsuan

Crowd counting is usually handled in a density map regression fashion, which is supervised via a L2 loss between the predicted density map and ground truth. To effectively regulate models, various improved L2 loss functions have been proposed to find

Externí odkaz: http://arxiv.org/abs/2212.04127

Zobrazit plný text záznamu

Report

Consistency-Aware Anchor Pyramid Network for Crowd Localization

Autor: Liu, Xinyan, Li, Guorong, Qi, Yuankai, Han, Zhenjun, Huang, Qingming, Yang, Ming-Hsuan, Sebe, Nicu

Crowd localization aims to predict the spatial position of humans in a crowd scenario. We observe that the performance of existing methods is challenged from two aspects: (i) ranking inconsistency between test and training phases; and (ii) fixed anch

Externí odkaz: http://arxiv.org/abs/2212.04067

Zobrazit plný text záznamu