Intelligent Scissors is an interactive tool for image segmentation and composition. It allows quick and accurate extraction of objects from digital images using simple mouse gestures. When the mouse approaches an object edge, a live-wire boundary "snaps" to and wraps around the object. Live-wire boundary detection uses discrete dynamic programming (DP) as a two-dimensional graph search problem, providing mathematically optimal boundaries while reducing sensitivity to local noise. Robustness is enhanced with on-the-fly training, which adapts the boundary to the specific edge being followed. Boundary cooling freezes unchanging segments and automates seed point input, allowing more freedom in gesture paths. Extracted objects can be scaled, rotated, and composited using live-wire masks and spatial frequency equivalencing. Frequency equivalencing is achieved by applying a Butterworth filter to match the lowest frequency spectra to other image components. Intelligent Scissors enable convincing compositions from existing images while increasing speed and precision in object extraction. The tool uses dynamic programming to find optimal paths between start and goal nodes, with boundary cooling and on-the-fly training reducing user input and adapting to specific edges. Live-wire masking and spatial frequency equivalencing allow for convincing image compositions. Compared to snakes, Intelligent Scissors offer more flexibility and efficiency. They allow interactive selection of optimal boundaries from all possible paths, rather than iterative refinement of a rough approximation. This makes them more efficient and user-friendly for image segmentation and composition. Intelligent Scissors can be applied to various image types, including color images, and can handle complex backgrounds. The tool uses subpixel edge filtering and anti-aliasing, multiple layered masks, and locally adaptive spatial frequency equivalencing. It also extends to moving objects and multiplane masks for image sequences. Results show that Intelligent Scissors are faster and more accurate than manual tracing. They can extract boundaries from a variety of images, including medical images, with efficient performance. The tool is intuitive and can be applied to existing black and white or color images of arbitrary complexity. Future work includes extending the tool to moving objects and multiplane masks for image sequences.Intelligent Scissors is an interactive tool for image segmentation and composition. It allows quick and accurate extraction of objects from digital images using simple mouse gestures. When the mouse approaches an object edge, a live-wire boundary "snaps" to and wraps around the object. Live-wire boundary detection uses discrete dynamic programming (DP) as a two-dimensional graph search problem, providing mathematically optimal boundaries while reducing sensitivity to local noise. Robustness is enhanced with on-the-fly training, which adapts the boundary to the specific edge being followed. Boundary cooling freezes unchanging segments and automates seed point input, allowing more freedom in gesture paths. Extracted objects can be scaled, rotated, and composited using live-wire masks and spatial frequency equivalencing. Frequency equivalencing is achieved by applying a Butterworth filter to match the lowest frequency spectra to other image components. Intelligent Scissors enable convincing compositions from existing images while increasing speed and precision in object extraction. The tool uses dynamic programming to find optimal paths between start and goal nodes, with boundary cooling and on-the-fly training reducing user input and adapting to specific edges. Live-wire masking and spatial frequency equivalencing allow for convincing image compositions. Compared to snakes, Intelligent Scissors offer more flexibility and efficiency. They allow interactive selection of optimal boundaries from all possible paths, rather than iterative refinement of a rough approximation. This makes them more efficient and user-friendly for image segmentation and composition. Intelligent Scissors can be applied to various image types, including color images, and can handle complex backgrounds. The tool uses subpixel edge filtering and anti-aliasing, multiple layered masks, and locally adaptive spatial frequency equivalencing. It also extends to moving objects and multiplane masks for image sequences. Results show that Intelligent Scissors are faster and more accurate than manual tracing. They can extract boundaries from a variety of images, including medical images, with efficient performance. The tool is intuitive and can be applied to existing black and white or color images of arbitrary complexity. Future work includes extending the tool to moving objects and multiplane masks for image sequences.