Visual Dataset of Traditional Portuguese Musical Instruments

This dataset was created to support classifying and preserving traditional Portuguese musical instruments using computer vision techniques. The hypothesis is that visual data when systematically organized and labelled, can effectively train machine learning models to recognize and classify Portuguese folk instruments. The dataset includes frames extracted from videos by 'A Música Portuguesa A Gostar Dela Própria' (MPAGDP), an association focused on documenting Portuguese cultural heritage. Five frames per video were selected and normalized to ensure consistency, enabling accurate analysis and model training. The data can be used to identify features of Portuguese instruments, supporting applications in digital ethnomusicology, automated classification, and cultural studies. Researchers can interpret the data as a tool for deep learning and comparative analysis of visual characteristics across instrument types. This dataset provides a structured foundation for studies in cultural preservation, deep learning, and visual pattern recognition.本数据集旨在借助计算机视觉（computer vision）技术，支撑葡萄牙传统乐器的分类与保护工作。本研究的核心假设为：经系统化组织与标注的视觉数据，可有效训练机器学习（machine learning）模型，实现葡萄牙民间乐器的识别与分类。该数据集包含由专注于葡萄牙文化遗产记录的协会“葡萄牙本土音乐爱好者协会（A Música Portuguesa A Gostar Dela Própria，简称MPAGDP）”从视频中提取的帧图像。研究人员为每份视频选取5帧并进行归一化处理，以保障数据一致性，为精准分析与模型训练提供支撑。 该数据集可用于识别葡萄牙乐器的视觉特征，为数字民族音乐学（digital ethnomusicology）、自动化分类以及文化研究等领域的应用提供支撑。研究人员可将该数据集作为深度学习（deep learning）工具，开展不同乐器类型视觉特征的对比分析研究。本数据集为文化遗产保护、深度学习以及视觉模式识别（visual pattern recognition）领域的研究提供了结构化的基础支撑。