This upload contains the raw projections of Sample 12-13, The data description and source are provided in the reference below: Zhou, E., Li, W., Xu, W. et al. A cone-beam photon-counting CT dataset for spectral image reconstruction and deep learning. Sci Data 12, 1955 (2025). https://doi.org/10.1038/s41597-025-06246-4 Abstract: Photon-counting CT has gained significant attention in recent years; however, publicly available datasets for spectral reconstruction and deep learning training remain limited. Consequently, many image process algorithms and deep learning models are developed and validated using simulated rather than real spectral CT data. To address this gap, we present a cone-beam photon-counting CT (PCCT) dataset acquired using a custom-built micro-PCCT system and 15 walnut samples. Each walnut was scanned from four bed positions under dual energy thresholds (15 keV and 30 keV), resulting in a total of 172,800 raw projection images with a resolution of 2063 × 505 pixels. The dataset provides full access to raw multi-energy projections, system parameters, calibration tables and reconstruction code, enabling comprehensive spectral CT studies including spectral CT reconstruction, material decomposition, artifact correction, and deep learning-based methods. It addresses the scarcity of real PCCT datasets for developing and validating data-driven approaches and aims to foster fair and reproducible comparisons across spectral CT image process algorithms. Methods: The dataset was acquired using a custom-built Micro photon-counting CT(Micro-PCCT),was jointly developed by Hainan University and United Imaging Life Science Instrument (LSI, Wuhan, China). The photon-counting detector has an effective resolution of 2063 × 505 pixels (after cropping peripheral invalid pixels), with a pixel size of 100 × 100 µm, and supports two independently adjustable energy thresholds. Each energy channel uses a 12-bit counter capable of recording up to 4096 photons per acquisition. Due to the limited detector width, each walnut was scanned at four axial bed positions along circular trajectories, spaced 15 mm apart, to ensure full coverage. The field of view (FOV) was set to 80 mm to encompass the entire walnut. Scans were performed in continuous mode, acquiring 1440 projections per circular trajectory with an angular increment of 0.25°.The imaging parameters are as follows: Para Value X-ray Source Tube voltage 80kV Tube current 200uA Filter 0.5mm AL Photon counting Detector Detector rows 2063 Detector columns 505 Detector pixel size 100μm Energy threshold 1 15kev Energy threshold 2 30kev Exposure time 70ms System setting Field of view 80mm Scan mode Continuous Number of projections per circle 1440 Source to object distance 140mm Source to detector distance 325mm Sample Number of Walnuts 15 Please refer to the paper for all further technical details The complete data set can be found via the following links: calibrationtable&sample1, sample2-3, sample4-5, sample6-7, sample8-9, sample10-11, sample12-13, sample14-15 The corresponding Matlab scripts for loading, pre-processing and spectral reconstructing the projection data in the way described in the paper can be found on github: https://github.com/zezisme/WalnutPCCTReconCodes For more information or guidance in using these dataset, please get in touch with: enzezhou@hust.edu.cn; tianwuxie@fudan.edu.cn