FMCG-RP2K-Shampoo-Dataset: A Fine-Grained Classification Dataset for Fast-Moving Consumer Goods

This dataset is a fine-grained classification dataset for fast-moving consumer goods (FMCG), focusing on shampoo product recognition. It contains product images, OCR results, text region importance scores, and multi-rule label matching scores, designed to support multimodal fine-grained classification research. The dataset includes: - Train and test splits of real-world shampoo product images - OCR recognition results in JSON format - Extracted text files from product packaging - Text region importance scores (area + position weighted) - Four-rule label matching scores for multimodal learning - Class index mapping file (class_indices.json) This dataset is associated with the paper submitted to The Visual Computer. If you use this dataset, please cite the corresponding paper.