Vessel enhancement with multiscale and curvilinear filter matching for placenta images

Placental vessel segmentation and registration in fetoscopy: Literature review and MICCAI FetReg2021 challenge findings

“…The work, developed and tested with ex-vivo images, combined Hessian-based filtering and a custom neural network trained on handcrafted features. The approach was improved by Chang et al (2013) , which introduced a vessel enhancement filter that combined multi-scale and curvilinear filter matching. The multi-scale filter extends the Hessian filter, introducing two scaling parameters to tune vesselness sensitivity.…”

“…Key: IFM - inter-fetus membrane; GMS - grid-based motion statistics; EMT - electromagnetic tracker. Reference Task Methodology Imaging type Almoussa et al (2011) Vessel segmentation Hessian filter and Neural Network trained on handcrafted features Ex-vivo Chang et al (2013) Vessel segmentation Combined Enhancement Filters Ex-vivo (150 images) Sadda et al (2019) Vessel segmentation Convolutional Neural Network (U-Net) In-vivo (345 frames from 10 TTTS procedures) Bano et al (2019) Vessel segmentation Convolutional Neural Network In-vivo (483 frames from 6 TTTS procedures) Casella et al (2020) IFM segmentation Adversarial Neural Network (ResNet) In-vivo (900 frames from 6 TTTS procedures) Casella et al (2021) IFM segmentation Spatio-temporal Adversarial Neural Network (3D DenseNet) In-vivo (2000 frames from 20 TTTS procedures) a Reeff et al (2006) Mosaicking Hybrid feature and intensity-based In water ex-vivo placenta Daga et al (2016) Mosaicking Feature-based with GPU for real time computation Ex-vivo, Phantom placenta Tella et al (2016) Mosaicking Combined EM and visual tracking probablistic model Ex-vivo w/laparoscope& EMT Gaisser et al (2016) M...…”

Analytical Approaches for the Segmentation of the Zebrafish Brain Vasculature

“…This is likely to be the case when working with vessels spanning a large size range where multiple scales need to be integrated. Further details are available elsewhere (Chang, Huynh, Vazquez, & Salafia, 2013).…”

Deep Placental Vessel Segmentation for Fetoscopic Mosaicking

“…Moreover, there are no publicly available annotated datasets to perform extensive supervised training. Methods based on a multi-scale vessel enhancement filter [14] have been developed for segmenting vasculature structures from ex vivo high-resolution photographs of the entire placental surface [1,11]. However, such methods fail on in vivo fetoscopy [22], where captured videos have significantly poorer visibility conditions, lower resolution, and a narrower FoV.…”