||Body dimensions are the basic information needed for the garment manufacturing industry. Traditionally the tailor measures all the dimensions directly on the human body. To collect a large number of human body dimension data, it will be a very tedious process requiring a lot of resources. With rapid development of the advanced electronics and optical hardware, intense research is conducted to devise a more efficient and sufficiently accurate method. Typical body measurement systems includes laser scanner, structure light and camera. They all have their pros and cons. This thesis presents a system based on camera imaging. Two pictures of the human body are taken by a CCD camera with a standard background. Body features are identified from the picture by a heuristic approach and body dimensions are derived from statistical fitting to a prevalent human anthropometric database. The algorithm will calculate the body dimensions automatically. Previous work has to rely on initial camera calibration. In this thesis, prescribed background geometrical information is utilized to help camera calibration. This makes it possible to conduct the body measurement in any arbitrary background. A statistical method is used to determine the relationship between feature distance in the 2D images and the 3D body dimensions from a prevalent human anthropometric database containing 1700 human data. The methodology estimates a total of 15 body dimensions based on feature data derived from the 2D imaging system.