The George B. Moody PhysioNet Challenges annually address critical health issues. The 2024 Challenge aims to digitize and categorize electrocardiograms (ECGs) from physical prints, enhancing cardiovascular disease understanding and treatment. Our team, Paper2Pulse, has two ideas to address this paper-to-digital signal conversion task. Firstly, a multi-stage modeling approach involves training an encoder-decoder-based noise-canceling model to clean ECG prints, followed by signal extraction. Secondly, an end-to-end modeling approach using pretrained networks like ViT directly outputs digitized signals. For the second classification task, ECG signal classification, we'll transform ECG signals into 2D features using techniques like log-mel spectrogram, Short Time Fourier Transform (STFT), and Constant Q Transform (CQT), and experiment with deep learning-based classification models.