capsule
发表于 2025-3-28 15:18:49
SP-FVR: SuperPoint-Based Finger Vein Recognitionthe SuperPoint model. Finally, the keypoints are matched by a bilateral strategy. Experiments on the HKPU and SDU-MLA databases demonstrate the effectiveness of the proposed method in real applications, the EERs are 0.0025 and 0.0138, with the FRRs-at-0-FAR of 0.0218 and 0.0890, respectively.
lattice
发表于 2025-3-28 20:16:53
http://reply.papertrans.cn/19/1882/188178/188178_42.png
不怕任性
发表于 2025-3-29 01:08:38
https://doi.org/10.1057/9780230502666 Among them, finger based multimodal biometric recognition is the most common and efficient way. However, biometric recognition has lots of problems such as too high feature dimensions, high computational complexity and insufficient correlation between modal classes. This paper propose an improved D
Boycott
发表于 2025-3-29 05:12:15
https://doi.org/10.1057/9780230502666k, which may affect the effectiveness of finger vein extraction. In this paper, we propose a global lightweight finger vein segmentation model, TRUnet, and build a lightweight Lightformer module and a plug-and-play module, Global-Lightweight block, in the proposed model respectively. The network not
Limited
发表于 2025-3-29 09:06:31
http://reply.papertrans.cn/19/1882/188178/188178_45.png
Parley
发表于 2025-3-29 13:14:19
http://reply.papertrans.cn/19/1882/188178/188178_46.png
蔓藤图饰
发表于 2025-3-29 15:45:13
http://reply.papertrans.cn/19/1882/188178/188178_47.png
Harrowing
发表于 2025-3-29 22:07:58
http://reply.papertrans.cn/19/1882/188178/188178_48.png
Respond
发表于 2025-3-30 01:22:22
http://reply.papertrans.cn/19/1882/188178/188178_49.png
senile-dementia
发表于 2025-3-30 07:43:12
http://reply.papertrans.cn/19/1882/188178/188178_50.png