摘要:
A Tree-Structured Decoder for Image-to-Markup Generation(https://www.zhuanzhi.ai/document/f09d28ca52b77a13a13bf44a86aba64d)Handwritten Mathematical Ex 阅读全文
摘要:
https://arxiv.org/pdf/1910.02562.pdf 总体介绍 基于Seq2Seq的OCR改进文章,提出两个问题,encoder特征间相似度太高导致注意力不准 & RNN-based的结构需要逐个step跑效率太低: 1、 such methods suffer from att 阅读全文
摘要:
2. Pre-processing Tricks 2.1. Dataset Cleaning for Training TR1 We noticed by visual inspection that the training set of GLD v1 is clean and reliable, 阅读全文
摘要:
对上采样的优化:动态的选取新像素对哪些原有像素存在依赖 Large field of view. Unlike previous works (e.g. bilinear interpolation) that only exploit subpixel neighborhood, CARAFE c 阅读全文