This work consists of the development of a computer scheme to provide the processing of digital mammographic images sent by an Internet user. The site, responsible for the interaction user/system, was developed in HTML and PHP with access to a data base implemented in MySQL. The routines of Pre-Processing, Segmentation, Classification and Generation of the Resultant Image were developed in Delphi and are executed parallel in the server. The current system results provide indications on the suspicious mammogram regions with the detected lesions. Besides the image with convenient marks on detected clustered microcalcifications, their classification in terms of “suspect” or “non-suspect” is also provided. The density level, as well as the percentage probabilities regarding the BIRADS® classification and the mass margin shapes are presented for suspicious masses detected by the scheme. In addition to the whole mammogram, the user can upload regions of interest (ROIs). Tests have shown that the correct rate of digitized mammograms was about 93% for microcalcification detection, while for mass detection, the correct rate was 92%. For direct digital mammograms, the correct rate was 93% for microcalcification detection, while for mass detection, it was 89%. In addiction, it was verified that the processing time average varied between 10 seconds (the best image case – one ROI) and 1,5 minutes for the worst image case (four complete mammograms), which this time can be considered acceptable. According to the tests performed with the purpose of checking the system efficacy, the tools manipulation was qualified as easy by more than 70% of the 25 volunteers whom have tested the system and its working classified as great and good by 40% and 56% of the users respectively. Currently, there are CAD schemes available on the market, however, they present a high acquisition cost and a final answer restricted just to the detection of suspect lesions, without providing additional data that can enhance the information the radiologists have, therefore helping them on their report. This research was carried out in order to provide this additional data by the Internet. Even though some problems occurred with the transmission of images by the Internet, the results presented by the tests performed by volunteers showed that the system has a good performance. This computational scheme can be accessed directly at http://143.107.235.167/CAD_Online/paginas/index.html or by the homepage of LAPIMO, the Laboratory for Medical and Odontological Images Analysis and Processing, at the Department of Electrical Engineering – EESC/USP.