Where do experts look while doing 3D image segmentation

Sanandaji, Anahita; Grimm, Cindy; West, Ruth; Parola, Max

doi:10.1145/2857491.2857538

Cited by 6 publications

(15 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The protocol and data analysis methods have yielded initial results in two areas: 1) where do experts look, what are their low-level actions, and how do gaze patterns change according to tasks during segmentation [22], and 2) what domain-agnostic high-level task patterns are exhibited by expert and novice segmenters. Together these represent progress towards elucidating low and high-level processes and constraints to address the hypotheses in Section 1.…”

Section: Initial Results and Hypothesesmentioning

confidence: 99%

“…Each action has a dedicated code. In a recent publication [22] we demonstrated that the protocol and coding scheme effectively captures, in a quantitative manner, the low-level tasks and behaviors and higher-level actions and constraints observed qualitatively during 3D volume segmentation by experts and novices. To obtain the initial results presented in Section 5 below we analyzed the data using the following methods: Action frequency analysis: We compute code frequency and average duration of each code interval over the entire observation.…”

Section: Discussionmentioning

confidence: 97%

“…Our subsequent analysis has focused on extracting commonalities that can be observed across these varying domains and tool sets, with the aim of separating tool and data-specific information from fundamental workflow. We have successfully identified both macro-work flow patterns (initial analysis, marking, evaluation) and microsub-task cycles [22] using eye tracking. Analysis of these workflows and mental models has led to a testable domain and tool agnostic hypothesis.…”

Section: Discussionmentioning

confidence: 99%

“…drawing) and high-level (e.g. marking versus reviewing actions) strategies in our recent publication [22].…”

Section: Interview Questions and Annotation Codesmentioning

confidence: 99%

“…Using the analysis of expert gaze and low-level actions described in Section 4 (above) we verified the data coding scheme can distinguish different patterns of low-level percep-tual actions, gaze patterns, and tasks. Detailed data analysis is presented in [22]. Key insights from these results include: Strategies: Depending on the tool and the data set, participants utilized one of two segmentation strategies: Regionbased (seeing data as regions to be filled or captured) or boundary-based (seeing data as boundaries to be demarcated by drawing contours).…”

Section: Perception and Low Level Actionsmentioning

confidence: 99%

See 4 more Smart Citations

Eliciting Tacit Expertise in 3D Volume Segmentation

West

Kajihara

Parola

et al. 2016

Proceedings of the 9th International Symposium on Visual Information Communication and Interaction

Self Cite

View full text Add to dashboard Cite

The output of 3D volume segmentation is crucial to a wide range of endeavors. Producing accurate segmentations often proves to be both inefficient and challenging, in part due to lack of imaging data quality (contrast and resolution), and because of ambiguity in the data that can only be resolved with higher-level knowledge of the structure and the context wherein it resides. Automatic and semi-automatic approaches are improving, but in many cases still fail or require substantial manual clean-up or intervention. Expert manual segmentation and review is therefore still the gold standard for many applications. Unfortunately, existing tools (both custom-made and commercial) are often designed based on the underlying algorithm, not the best method for expressing higher-level intention. Our goal is to analyze manual (or semi-automatic) segmentation to gain a better understanding of both low-level (perceptual tasks and actions) and high-level decision making. This can be used to produce segmentation tools that are more accurate, efficient, and easier to use. Questioning or observation alone is insufficient to capture this information, so we utilize a hybrid capture protocol that blends observation, surveys, and eye tracking. We then developed, and validated, data coding schemes capable of discerning low-level actions and overall task structures.

show abstract

Section: Initial Results and Hypothesesmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 97%

Section: Discussionmentioning

confidence: 99%

“…drawing) and high-level (e.g. marking versus reviewing actions) strategies in our recent publication [22].…”

Section: Interview Questions and Annotation Codesmentioning

confidence: 99%

Section: Perception and Low Level Actionsmentioning

confidence: 99%