用边缘梯度训练全连接网络，轻松达到手写字符识别高精度。
A Sobel-Gradient MLP Baseline for Handwritten Character Recognition
- 仅用Sobel算子提取的水平与垂直梯度作为输入
- 在MNIST上达98%准确率，EMNIST字母集达92%
- 模型轻量透明，适合对可解释性要求高的场景
我们重新审视经典的Sobel算子，提出一个简单问题：仅用一阶边缘图能否驱动全连接多层感知机（MLP）完成手写字符识别（HCR），替代卷积神经网络（CNN）？仅以水平和垂直Sobel导数作为输入，在MNIST和EMNIST Letters数据集上训练MLP。尽管结构极简，该模型在MNIST数字识别上达到98%准确率，在EMNIST字母识别上达92%，接近CNN表现，同时具备更小内存占用和更透明特征表示。结果表明，手写字符图像中的大部分类别判别信息已由一阶梯度充分捕获，使基于边缘感知的MLP成为HCR的有力候选方案。
原文摘要 · Abstract (English)
We revisit the classical Sobel operator to ask a simple question: Are first-order edge maps sufficient to drive an all-dense multilayer perceptron (MLP) for handwritten character recognition (HCR), as an alternative to convolutional neural networks (CNNs)? Using only horizontal and vertical Sobel derivatives as input, we train an MLP on MNIST and EMNIST Letters. Despite its extreme simplicity, the resulting network reaches 98% accuracy on MNIST digits and 92% on EMNIST letters -- approaching CNNs while offering a smaller memory footprint and transparent features. Our findings highlight that much of the class-discriminative information in handwritten character images is already captured by first-order gradients, making edge-aware MLPs a compelling option for HCR.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。