Mattysmittttt/camonet-synthetic

CamoNet Synthetic是一个程序生成的军事伪装图案数据集，包含40种历史和当代的伪装图案，每种图案有200个样本，总计8,000张256×256的RGB图像。每张图像都标注了来源、时代和视觉家族。数据集完全由合成图像组成，通过小型Python脚本生成，确保了数据的可重复性和自由再分发性。数据集适用于图像分类任务，主要用于增强现实照片数据集、预训练视觉模型、调试管道和进行类别平衡实验。然而，它不适合用于生产级OSINT分类器的独立训练或需要色彩保真度的法医/法律图案归属工作。

CamoNet Synthetic is a procedurally-generated military camouflage pattern dataset containing 40 historical and contemporary patterns with 200 samples each, totaling 8,000 256×256 RGB images. Each image is tagged with origin, era, and visual family. The dataset is fully synthetic, generated from small Python recipes per pattern family, ensuring reproducibility and free redistribution. It is suitable for image classification tasks, primarily for augmenting real photo datasets, pre-training vision models, debugging pipelines, and conducting class-balance experiments. However, it is not suitable for standalone training of production OSINT classifiers or forensic/legal pattern attribution work requiring colorimetric fidelity.