scikit-image

scikit-image (skimage) is an open-source image processing library for Python, built on NumPy and SciPy. Part of the scientific Python ecosystem alongside NumPy, SciPy, Matplotlib, and Pandas, scikit-image provides algorithms for image processing and computer vision tasks. Maintained by Stefan van der Walt, Juan Nunez-Iglesias, and the scientific Python community, scikit-image has over 6,200 stars as of 2026. Unlike general-purpose computer vision libraries like OpenCV (which focuses on real-time applications and C++ performance), scikit-image is designed for scientific image analysis with an emphasis on algorithm correctness, documentation, and Pythonic API design. Key features include: image I/O (reading and writing various formats including TIFF, PNG, JPEG via the io module), color space conversion (RGB, HSV, LAB, XYZ, grayscale, and perceptual color spaces), geometric transformations (resizing, rotation, warping, affine/perspective transforms with anti-aliasing), feature extraction (corner detection via Harris and Shi-Tomasi, blob detection, ORB feature detection and matching, template matching), segmentation (watershed, active contours, random walker, SLIC superpixels, Chan-Vese), morphology operations (erosion, dilation, skeletonization, convex hull), filtering (Gaussian, median, Wiener, bilateral, non-local means denoising), restoration (Richardson-Lucy deconvolution, inpainting), measure (region properties, contour finding, structural similarity index SSIM), and graph-based image analysis.