Introduction There is almost no area of technical endeavor that is not impacted in some way by digital image processing . The revolutionary discovery of the Fast Fourier Transform (FFT) algorithm has allowed engineers and scientists to apply the powerful techniques of Fourier analysis to digital signals at vastly reduced computational complexity, spurring explosive growth in the field of image processing. The continual advances in transistor and VLSI technologies have led to rapid improvements in microprocessor performance so that even complex image processing algorithms can now be trivially performed by personal computers. This paper explores the mathematical models and algorithms developed within the domain of digital image processing for the purpose of autonomous object recognition and motion analysis. Commercial Applications Object recognition and motion analysis technologies both see application in a wide number of commercial products. Many modern digital cameras come equipped with auto-focus and face detection systems . Some sophisticated video surveillance systems feature autonomous detection of unauthorized entry through the use of face recognition software. Motion analysis algorithms also commonly see use in surveillance systems, e.g., in the automatic tracking of motion for security purposes. Quite often the two tasks are employed simultaneously, as in automatic surveillance of elderly patients in hospitals. If an object is recognized to be human (e.g., a patient) and if the motion of the object is determined to be potentially dangerous (e.g., a fall), then the appropriate persons are notified. Underlying Technology Most object recognition algorithms involve the use of edge detection techniques. Some approaches attempt to isolate the object for analysis by extraction through foreground segmentation . Other algorithms seek to model the object of interest as an ellipse (or other shape), so as to analyze the degree of statistical similarity to an ideal candidate object , . Optical flow techniques are commonly used to detect motion in a video sequence. However, optical flow is not well-suited for real-time video analysis and is particularly susceptible to noise . Alternatively, some motion analysis algorithms calculate and utilize Motion History Images (MHI), where pixel intensities correspond to the recency of motion in an image sequence . MHI-based algorithms generally offer improved real-time performance over those based on optical flow. Implementation Object recognition and motion analysis are generally implemented via software written in high level languages suitable for digital signal processing, e.g. C/C++ and MATLAB. A video camera capable of interfacing with a computer or embedded processor is required for data acquisition. The method of interfacing must allow for the image data to be written to memory faster than the data is obtained (i.e., the camera frame rate). A common implementation involves a wide-angle webcam interfaced to a PC via USB . Sources  R. C. Gonzalez and R. E. Woods. Digital Image Processing, 3rd edition. Upper Saddle River, New Jersey: Pearson Prentice Hall, 2008.  Canon USA. PowerShot SD1000 Technical Specifications. PowerShot SD1000 Digital Camera, Feb. 21, 2007. [Online]. Available: http://www.usa.canon.com/consumer/controller?act=ModelInfoAct&fcategoryid= 145&modelid=14901#ModelTechSpecsAct. [Accessed: Jan. 20, 2009].  K. Kim, T. Chalidabhongse, D. Harwood, and L. Davis. Real-time foregroundbackground segmentation using code-book model. Real-Time Imaging, 11(3):172185, June 2005.  C. Rougier, J. Meunier, A. Arnaud, and J. Rousseau. Fall detection from human shape and motion history using video surveillance. In IEEE Advanced Information Networking and Applications Workshops, 2007.  W. Pratt. Digital Image Processing, 3rd edition. New York: John Wiley & Sons, 2001.  A. Bobick and J. Davis. The recognition of human movement using temporal templates. In IEEE Transactions on Pattern Analysis and Machine Intelligence, volume 23, pp. 257-267, March 2001.