training images چیست؟

[M3RS4D 50062]

بسم الله الرحمن الرحیم
با سلام
قسمتی از مقاله که مرتبط با سوال هست براتون گذاشتم:

We assume we are given a set of segmented training images,
in which the boundaries between the segments correspond
to contours [1, 33]. Given an image patch, its annotation
can be specified either as a segmentation mask indicating
segment membership for each pixel (defined up to a
permutation) or a binary edge map. We use y ∈ Y = Zd×d
to denote the former and y ∈ Y = {0, 1}d×d for the latter,
where d indicates patch width. An edge map y can always
be trivially derived from segmentation mask y, but not vice
versa. We utilize both representations in our approach.
Next, we describe how we compute the input features x,
the mapping functions Πφ used to determine splits, and the
ensemble model used to combine multiple predictions.

با سپاس فراوان

یه توضیحی در موردش میشه بدین؟!؟!؟
من متوجه نشدم !