PSI - Issue 84

Regina Finocchiaro et al. / Procedia Structural Integrity 84 (2026) 1264–1269

1267

A third node emerges in the translation from qualitative description to both Class of Attention (CdA) labelled assignment and consequent selection of further investigations. When inputs are qualitative or insufficiently comparable due to different campaign approaches, the perceived vulnerability may shift not only because the asset is really different, but because the observation has been framed, grouped or documented differently. Furthermore, observational uncertainty associated with incomplete coverage, limited accessibility, or unfavourable environmental conditions can remain implicit or “invisible” in the dataset, even though it is decisive in determining whether L3/L4 investigations are warranted and how they should be targeted. 5. A “soft” integration of Gestalt: Gestalt-Informed Bridge Guideline Augmentation (GIBGA) proposal In this paper the proposed Bridge Guideline integration is conceived as a reinforcement of the information workflow, not as a replacement of forms, classes or Code criteria. The aim is to add traceability to cognitive transition that transforms signs into phenomena and phenomena into decisions. In summary, the approach relies on four operational principles that must be implemented at an operator’s mandatory request: • Transparency of grouping: a concise declaration of the criteria used to aggregate signs has to be defined. • Separation between observed and inferred: distinguishing what is directly seen (signs) from what is interpreted (phenomenon/mechanism) has to be carried out in order to preserve process auditability. • Document traceability: improving localization, comparability and reuse of images and operator annotations over time (subsequent campaigns, handovers, audits) have to be guaranteed. • Explicit management of uncertainty: when interpretation is ambiguous or data coverage is incomplete, the uncertainty declaration must become useful information rather than an implicit detail. Within this framework, concise perceptual metadata can be introduced to qualify, in a lightweight but helpful manner, both the survey quality and interpretation robustness (e.g. according to definitions of SACK acronym proposed by the Authors). The purpose is not to define a determinate “measure perception” but to create a minimal layer of declared perceptual quality that becomes useful when the procedure growing must decide survey priorities and further investigations. Table 1. Table between Gestalt principles, inspection-related risks and operational countermeasures proposed within the Authors’ Gestalt Informed Bridge Guideline Augmentation (GIBGA) framework. Gestalt principle What mind tends to do Inspection risk “GIBGA” countermeasure

Groups elements that are close to each other

Merges independent defects into a single phenomenon

Make the grouping criterion explicit; verify a common cause; make reference to an element grid/structural breakdown Separate sign vs phenomenon; use multi scale photos and comparison with a defined defect catalogue

Proximity

Groups by similarity (shape/texture/colour/orientation)

Over-extends a pattern (e.g., superficial spots and marks interpreted as corrosion phenomena) Assumes crack/lesion continuity beyond actual discontinuities

Similarity

Perceives aligned/curved segments as a single continuous trace

Mandatory segmentation; mark discontinuities; include measurements/scale

Good Continuation

Completes missing contours/parts

“Completes” defects hidden by repairs/occlusions

Controlled closure as a recording of defect pattern inference; trigger confirmation (via further NDT or special access) Checklist for low-salience zones; standardized photo set; two-pass scan (global + targeted) Record alternative hypotheses when ambiguity is high; provide inter-inspector audit/peer review

Closure

Selects what stands out as “figure” and suppresses background

Neglects diffuse, low-salience deterioration

Figure– Ground

Prefers simple, stable interpretations

Over-simplified diagnosis

Prägnanz (Good Form)