canonical
发表于 2025-3-26 22:57:36
Yuxin Du,Jing Fan,Ari Happonen,Dassan Paulraj,Micheal Tuapend approaches that directly adopt OCR features as the input of an information extraction module, we propose to use contrastive learning to narrow the semantic gap caused by the difference between the tasks of OCR and information extraction. We evaluate the existing end-to-end methods for VIE on the
CORD
发表于 2025-3-27 03:18:41
Tobias Dorrn,Almuth Müllerve experiments show that the proposed method outperforms the existing two-stage cascade models and one-stage end-to-end models with a lighter and faster architecture. Furthermore, the ablation studies verify the generalization of our method, where the proposed modal adapter is effective to bridge va
erythema
发表于 2025-3-27 05:40:27
Wisam Bukaita,Guillermo Garcia de Celis,Manaswi Gurram enhance recognition. Experiments on three datasets prove our method can achieve state-of-the-art recognition performance, and cross-dataset experiments on two datasets verify the generality of our method. Moreover, our method can achieve a breakneck inference speed of 104 FPS with a small backbone
间谍活动
发表于 2025-3-27 11:13:23
Yeferson Torres Berru,Santiago Jimenez,Lander Chicaiza,Viviana Espinoza Loayzar proposed approach outperforms several existing state-of-the-art approaches, including complex approaches utilizing generative adversarial networks (GANs) and variational auto-encoders (VAEs), on 7 of the datasets, while achieving comparable performance on the remaining 2 datasets. Our findings sug
AMITY
发表于 2025-3-27 14:10:27
Xiaoting Huang,Xuelian Xi,Siqi Wang,Zahra Sadeghi,Asif Samir,Stan Matwined on general domain document images, by fine-tuning them on an in-domain annotated subset of EEBO. In experiments, we find that an appropriately trained image-only classifier performs as well or better than text-based poetry classifiers on human transcribed text, and far surpasses the performance o
STALE
发表于 2025-3-27 19:37:48
Dorsa Soleymani,Mahsa Mousavi Diva,Lovelyn Uzoma Ozougwu,Riasat Mahbub,Zahra Sadeghi,Asif Samir,Stan Matwine-of-the-art in both datasets, achieving a word recognition rate of . and a 2.41 DTW on IRONOFF and an expression recognition rate of . and a DTW of 13.93 on CROHME 2019. This work constitutes an important milestone toward full offline document conversion to online.