Two-phase flow in channels of small dimensions is often a non-stationary process, the nature of such flow is oscillatory. Due to small channel dimensions, high heat flux, parallel channels interactions, pressure and temperature oscillations, the character of the phenomena occurring during boiling is complex. The changes of the measured signals are observed in different time scales. In order to examine in detail two-phase flow parameters changes, many acquisition devices are often installed. This solution becomes challenging concerning mini and microchannel heat-exchangers due to space limitation and modifications of an experimental setup. This paper presents a novel application of multiscale entropies for spatial and temporal analysis of two-phase flow based on only one registered parameter. This analysis is performed based on pixel brightness changes in photo frames registered by a high speed camera during two-phase flow. The spatial changes of pixel brightness are observed on single frames and temporal changes are examined using a set of frames (in time). The Composite Multiscale Sample Entropy is applied to identify two-phase flow patterns and to analyze the complexity of phase distribution. Using Multivariate Mul-tiscale Sample Entropy the most rapid changes of phase distribution in a multichannel heat exchanger are determined.