Výsledky vyhledávání

Report

MetaSeg: MetaFormer-based Global Contexts-aware Network for Efficient Semantic Segmentation

Autor: Kang, Beoungwoo, Moon, Seunghun, Cho, Yubin, Yu, Hyunwoo, Kang, Suk-Ju

Beyond the Transformer, it is important to explore how to exploit the capacity of the MetaFormer, an architecture that is fundamental to the performance improvements of the Transformer. Previous studies have exploited it only for the backbone network

Externí odkaz: http://arxiv.org/abs/2408.07576

Zobrazit plný text záznamu

Report

Cross-aware Early Fusion with Stage-divided Vision and Language Transformer Encoders for Referring Image Segmentation

Autor: Cho, Yubin, Yu, Hyunwoo, Kang, Suk-ju

Referring segmentation aims to segment a target object related to a natural language expression. Key challenges of this task are understanding the meaning of complex and ambiguous language expressions and determining the relevant regions in the image

Externí odkaz: http://arxiv.org/abs/2408.07539

Zobrazit plný text záznamu

Report

AttentionHand: Text-driven Controllable Hand Image Generation for 3D Hand Reconstruction in the Wild

Autor: Park, Junho, Kong, Kyeongbo, Kang, Suk-Ju

Recently, there has been a significant amount of research conducted on 3D hand reconstruction to use various forms of human-computer interaction. However, 3D hand reconstruction in the wild is challenging due to extreme lack of in-the-wild 3D hand da

Externí odkaz: http://arxiv.org/abs/2407.18034

Zobrazit plný text záznamu

Report

Embedding-Free Transformer with Inference Spatial Reduction for Efficient Semantic Segmentation

Autor: Yu, Hyunwoo, Cho, Yubin, Kang, Beoungwoo, Moon, Seunghun, Kong, Kyeongbo, Kang, Suk-Ju

We present an Encoder-Decoder Attention Transformer, EDAFormer, which consists of the Embedding-Free Transformer (EFT) encoder and the all-attention decoder leveraging our Embedding-Free Attention (EFA) structure. The proposed EFA is a novel global c

Externí odkaz: http://arxiv.org/abs/2407.17261

Zobrazit plný text záznamu