Robust Segmentation Based Tracing Using an Adaptive Wrapper for Inducing Priors

Abstract

Abstract-Segmentation based tracing algorithms detect the extent and borders of an object in a given frame I Z by propagating results from frames I 1≤z<Z . Although application specific tracers have been forthcoming, techniques that automatically adapt across applications have been less explored. We approach this problem by learning a prior model on topological dynamics that encourages segmentation transitions across frames that are most likely for a given application. Further, we augment a generic tracing technique with a locality sensitive prior derived from dense optic flow fields for deformation guidance. The proposed approach comprises two stages where the generic tracer initially yields multiple segmentation transitions when its parameters are perturbed, and the learnt topology prior subsequently propagates high scoring segmentations. Because the learnt topology model wraps around a generic tracer and adapts it by setting its free parameters, the need for careful parameter tuning is completely obviated. Through extensive experimental validation in surveillance, biological and medical image datasets, we verify the applicability of the proposed model while demonstrating good tracing performance under severe clutter.Index Terms-Tracing, parameter adaptation, electron micrograph, Markov random fields.

High-Resolution Synaptic Connectomics

“…a weaker concept is the gold standard, which is the best reference that can be achieved under certain limitations. arguably, once 2-nm datasets exist with appropriate instances of 0.25-nm validation [3,43], lower resolution gold standard sets are of limited value.…”

“…even though it contains over 300 bipolar cells (Bcs), 300 Müller cells (Mcs), 39 aii acs, over 100 acs, and 20 ganglion cells (gcs), and includes a full set of small molecule markers for classification, it is far from complete. a spectrum of solutions has been proposed to complete the annotation set, ranging from automated detection schemas [43] to high-density crowdsourcing [15,38). We were one of the earliest proponents of crowdsourcing [2] and have retreated from that position somewhat to advocate an executive annotator + team format.…”

“…The ease with which both software and novice annotators confuse oblique synapses, adherens junctions, and gap junctions, demands a ready-correction schema. in theory, any automated system must be tested against a ground truth dataset [43], but no one has proven that any automated system can discriminate these components. We have three practical solutions.…”

2D hidden Markov model with spatially adaptive state-space for tracing many cells in image sequence

“…Other than superpixels or supervoxel-based grouping methods, in [14] and [15], a set of over-segmentation results are generated as multiple hypothesis, and graph cuts are utilized for optimizing the merging results. In this paper, we propose a pixel-wise labeling framework, 2D-HMM-SASS, to effectively and efficiently solve the problem of cell tracing on a highly challenging EM image sequence of a rabbit retina.…”

A grouping-feature and Nesting-Kernel scene image segmentation algorithm

“…However, the results may exist a very large number of small and quasi-homogeneous regions. Edge-based technique such as edge detectors [6], border tracing [7], and active contours [8], are based on the changes in the image edges. However, the results are also not so good in the case many edges in the images.…”