Fig. 1. Schematic of dynamic speckle OCTA system

Fig. 2. Flow chart of CEF algorithm

Fig. 3. Top view of dynamic speckle OCTA expanded into a series of eigenimages

Fig. 4. Top view of dynamic speckle OCTA expanded into first eigenimage and high-pass eigenimage. (a) Original top view; (b) first eigenimage; (c) high-pass eigenimage

Fig. 5. First eigenimage expanded into remodeled first eigenimage and orthogonal compensated product. (a)First eigenimage; (b) corresponding left singular vector and remodeled left singular vector; (c) right singular transpose vector; (d) remodeled first eigenimage; (e) orthogonal compensated product

Fig. 6. Comparison of results of destriping image processed by different algorithms. (a) Traditional eigenimage filtering algorithm; (b)CEF algorithm; (c) frequency rejection filtering algorithm; (d) tensor voting approach algorithm

Fig. 7. Relevant results of mouse experiment. (a) Original top view ; (b)first eigenimage; (c) remodeled first eigenimage; (d) orthogonal compensatory product; (e) high-pass eigenimage; (f) image processed by CEF algorithm; (g) image processed by frequency rejection filtering; (h) image processed by tensor voting approach algorithm