Imposing shift-invariance using Flexible Structure Dictionary Learning (FSDL)

Signal Processing for Newborn Survival from labour to resuscitation

“…In this work we propose to use shift-invariant dictionary by utilizing SI-FSDL, a dictionary learning method for shift invariant dictionaries recently proposed by our group [80]. SI-FSDL is capable of handling variable shifts and length of the atoms.…”

“…For imposing the shift-invariant structure onto the dictionary we utilized our previously proposed method, SI-FSDL [80]. One of the benefits of shift-invariant dictionaries is that we can address larger gaps (by using larger shift-invariant atoms) than a general dictionary while keeping the number of free variables fixed.…”

On learning with shift-invariant structures

“…Previously, several dictionary learning techniques that accommodate for shift invariance have been proposed: extending the well-known K-SVD algorithm to deal with shift-invariant structures [17], [20], [21], proposing a shift-invariant iterative least squares dictionary learning algorithm [22], extending the dictionary while solving an eigenvalue problem [23], fast online learning approach [24], research that combines shift and 2D rotation invariance [25] and proposing new algorithms that optimize directly the dictionary learning objective functions with circulant matrices [14], [15], [26]. The convolutional sparse representation model [27], [28], [29] where the dictionary is a concatenation of circulant matrices has been extensively studied in the past.…”