Slope difference distribution (SDD) is computed from the one-dimensional curve and makes it possible to find derivatives that do not exist in the original curve. It is not only robust to calculate the threshold point to separate the curve logically, but also robust to calculate the center of each part of the separated curve. SDD has been used in image segmentation and it outperforms all classical and state of the art image segmentation methods. SDD is also very useful in calculating the features for pattern recognition and object detection. For the gesture recognition, SDD achieved 100% accuracy for two public datasets: the NUS dataset and the near-infrared dataset. For the object recognition, SDD achieved 100% accuracy for the Kimia 99 dataset. In this memorandum, I will demonstrate the effectiveness of SDD with some typical examples.