Medical Image Cropping and 2.5D Slicing Generation Software

Development Background In the research of medical image deep learning, the raw data often contains a large number of irrelevant background regions, and directly inputting them into the model will reduce efficiency and introduce noise. Meanwhile, to adapt to 2D convolutional neural networks, it is often necessary to extract key slices from 3D volumes. Existing tools rely heavily on programming scripts, lack graphical interfaces, and are difficult to ensure slice direction consistency and visual quality. In addition, the window width and level settings of CT images are crucial for lesion visualization, but general image viewers cannot retain such medically specific information. To this end, we have developed this software that integrates three major functions: mask guided ROI cropping, anatomical direction standardization, and 2.5D slice generation with configurable window width and position into a single graphical interface, providing a "one click" intelligent preprocessing experience and significantly reducing the technical threshold. Software Purpose This software is designed specifically for medical image preprocessing and is used to perform the following operations on paired NIfTI (. nii. gz) format images and masks: 1) Intelligent ROI clipping: automatically locates lesion areas based on masks, supports isotropic cube output; 2) Directional standardization: redirects images to the standard RAS+anatomical coordinate system to ensure spatial consistency; 3) 2.5D slice generation: Generate JPG slices with correct orientation and uniform resolution (224 × 224) based on the mask center layer; 4) Medical visualization optimization: supports configuring CT window width and position to improve the visibility of lesions in JPG slices; 5) Batch processing: Automatically traverse all pairs of. nii.gz files with the same name in the input directory, suitable for preprocessing large-scale datasets. Software functions and features 1) Zero environment dependency: No need to install Python or any third-party libraries, just unzip and use; 2) Dual mode output: a. Crop mode: Output standardized NIfTI files while retaining the original intensity values; b. 2.5D slicing mode: outputting JPG image sequences with direction correction and window width optimization; 3) Intelligent center positioning: symmetrically extract multiple layers of slices centered on the maximum area layer of the mask, and accurately focus on the lesion; 4) Medical display optimization: supports custom CT window width and position (such as soft tissue window, lung window), JPG contrast is more in line with clinical habits; 5) Security priority: forced separation of input/output directories to prevent accidental overwriting of raw data; 6) Graphic interface: Designed with a dark theme, the operation is simple and intuitive, making it easy for non-technical users to get started.