Microscopy is of high interest for biology since it allows imaging features that are too small to be seen with naked eyes. However, cells are mostly transparent to visible and infrared light which makes it difficult to see with a traditional microscope. To do so, phase imaging was invented by F. Zernike, who received the Nobel prize in 1953 for this invention. It provides intensity contrast to visualize transparent samples such as those found in biology without any staining. Among the different implementations of phase imaging, digital holographic microscopy (DHM) is a well-known phase method which provides a quantitative value of the phase generated by the sample under test. Implementations of DHMs without the help of any lens element (so called lensless) offer the added advantage to provide large field of views (severalmm2 compared to several hundred ¹m2) and more compact setups than traditional DHMs which have high quality microscope objectives. The lateral resolution is limited by the pixel size of the camera. However, several techniques have been developed to obtain subpixel resolution with lensless setups, hence giving a resolution smaller than the pixel size. In this thesis, two lensless transmission DHMs are presented using a side illumination technique in order to further reduce the device size and keep a visual access to the sample. The first one operates in an in-line configuration: the most compact but using time-consuming image post-processing. The second one is also lensless and uses an off-axis configuration allowing quasi-real time imaging but less compact. Their practical use is described and images of transparent (phase only) samples are shown. Super-resolution (subpixel) using several subpixel shifted holograms was then investigated for the in-line configuration and demonstrated on absorptive samples. Since those microscopes are compact and made out of off-the-shelf low priced elements, they could be broadly spread in schools for educational purposes. They could also be implemented in cells incubators, allowing visualization inside processes at low cost and multiplication of the number of measurements made at the same time.