Time is a parameter playing a central role in our most fundamental modelling of natural laws. Relativity theory shows that the comparison of times measured by different clocks depends on their relative motion and on the strength of the gravitational field in which they are embedded. In standard cosmology, the time parameter is the one measured by fundamental clocks (i.e., clocks at rest with respect to the expanding space). This proper time is assumed to flow at a constant rate throughout the whole history of the universe. We make the alternative hypothesis that the rate at which the cosmological time flows depends on the dynamical state of the universe. In thermodynamics, the arrow of time is strongly related to the second law, which states that the entropy of an isolated system will always increase with time or, at best, stay constant. Hence, we assume that the time measured by fundamental clocks is proportional to the entropy of the region of the universe that is causally connected to them. Under that simple assumption, we find it possible to build toy cosmological models that present an acceleration of their expansion without any need for dark energy while being spatially closed and finite, avoiding the need to deal with infinite values.

Strong gravitational lensing gives access to the total mass distribution of galaxies. It can unveil a great deal of information about the lenses dark matter content when combined with the study of the lenses light profile. However, gravitational lensing galaxies, by definition, appear surrounded by point-like and diffuse lensed signal that is irrelevant to the lens flux. Therefore, the observer is most often restricted to studying the innermost portions of the galaxy, where classical fitting methods show some instabilities. We aim at subtracting that lensed signal and at characterising some lenses light profile by computing their shape parameters. Our objective is to evaluate the total integrated flux in an aperture the size of the Einstein ring in order to obtain a robust estimate of the quantity of ordinary matter in each system. We are expanding the work we started in a previous paper that consisted in subtracting point-like lensed images and in independently measuring each shape parameter. We improve it by designing a subtraction of the diffuse lensed signal, based only on one simple hypothesis of symmetry. This extra step improves our study of the shape parameters and we refine it even more by upgrading our half-light radius measurement. We also calculate the impact of our specific image processing on the error bars. The diffuse lensed signal subtraction makes it possible to study a larger portion of relevant galactic flux, as the radius of the fitting region increases by on average 17\%. We retrieve new half-light radii values that are on average 11\% smaller than in our previous work, although the uncertainties overlap in most cases. This shows that not taking the diffuse lensed signal into account may lead to a significant overestimate of the half-light radius. We are also able to measure the flux within the Einstein radius and to compute secure error bars to all of our results.