SCIENTIFIC ABSTRACT: IRAS has opened a completely new field in showing that active galaxies can be very powerful in the far-IR (IRAS galaxies) with luminosities as high or higher than 10^12 times the solar luminosity. There is also increasing evidence that the most violent bursts of star formation and the diverse manifestations of nuclear activity in galaxies are triggered by mergers. Although the gas distribution and starburst regions become more centrally concentrated as the merger evolves, it is known that the bursts of star formation are not only confined to the nuclei. A burst of star formation can also result from the existence of a stellar bar, through the combined action of a bar and a gravitational interaction. In all these galaxies, the molecular gas appears to be very concentrated, but even more so in the case of an active nucleus. While it is probable that the IR emission is due to dust thermal emission, the exact heating mechanism is not clear and can be actually multiple - heating by newborn stars or by the non-thermal radiation of an active nucleus - and will change with the distance to the nucleus. In this field, only little mapping has been done in the IR, but mainly spectroscopy, and it is obvious that much will be learned by mapping in the continuum (dust emission) and in selected lines: fine-structure lines for the ionized regions in a large range of temperatures, molecular hydrogen and fine-structure lines for shocks, IR bands for PAHs, etc. The spatial distribution of both dust and gas emission will be observed with ISOCAM which offers a high and homogeneous sensitivity, and a good angular resolution. We propose to observe a sample of galaxies in which the activity is due to either a high level of massive star formation (starburst galaxies) or an active nucleus (ex. Seyfert galaxies, quasars) or both. OBSERVATION SUMMARY: The list of targets consists in a sample of characteristic objects representative of galaxy mergers, nearby starbursts, nearby active galaxies, ultraluminous infrared galaxies which represents the high end of the merger class, and radio galaxies and quasars. Due to different brightness levels, several observational strategies are applied: - complete CVF coverage for the brightest (and closest) objects: NGC 253, NGC 1068, M82, Mrk 171, NGC 5128, NGC 3256 and NGC 1808. - a 150"x150" raster map of the Antennae (NGC 4038/9) with the LW2, LW3, LW4 and LW7. In the case of a spring launch, a 200"x200" raster map of NGC 1808 will be performed in the LW2 and LW3 filters. - imaging in several broad filters (LW2, LW3) and narrow filters LW4 (PAH) and LW7 (silicates) for the intermediate flux objects (objects with a dedicated time of 1320s in the list) - imaging only in broad filters (LW2 and LW3) for the most distant objects. As the interesting field is generally small, we will use 3 arcsec pixels with the idea that re-centering on probable bright spots will allow to use the full angular resolution. Due to its large angular extent, NGC 253 will be imaged with the 6 arcsec PFOV.