A histogram is a representation of the distribution of data in an image, e.g., gray-levels or colors. This common tool is used for many applications and especially for classification purposes. This key-feature has to be generically implemented in the Milena library. In this seminar, we propose to store histogram data using an image container. To that aim, adapting the definition of value types is required. More specifically, we propose to augment the traits associated with value types, add some new useful value types, and design new abstractions over them. Last, we present how to deal with circular data, such as the "hue" values (encoded by angles in the HSL color space); in this case, the "histogram image" would become also circular.