Abstract
The networks have been developing strongly these last years and became inevitable<br />for the modern communication. The images transmitted on these networks are particular<br />data because of their large quantity of information. The transmission of the images<br />thus raises a significant number of problems which the majority of them have no solution<br />yet. We enumerate for example safety, confidentiality, integrity and the authenticity of<br />these data during their transmission. Some medical applications require a strict association<br />between the image and its contextual data. The protection of high resolution<br />information (high frequencies of the images, details, scalable visualization) currently<br />has a high demand. During this thesis, our research led to the creation of three new<br />methods to make safe the transfer of images. The first two methods are based on hybrid<br />codings : use of encryption, data hidden and compression. The third approach is based<br />on the work of Droogenbroeck and Benedett. We propose selective encryption to protect<br />the transmission of images. It allows to selectively encrypt the image by protecting information<br />from high frequencies while remaining compatible according to the standards<br />of image compression.