Primerjava zaznavanja patologije na CT slikah glave in toraksa v različnih barvnih paletah : diplomsko delo

Kelc Hočevar, Pia (author), Kožuh, Rebeka (author), Žibert, Janez (mentor), Kaučič, Aleš (co-mentor), Alukić, Erna (reviewer)

Uvod: Človeško oko zazna veliko več odtenkov barv (približno milijon) kakor sivin (približno 30). Različne barve na posameznika različno vplivajo zaradi čustvenih in kognitivnih asociacij. Slike z različnimi algoritmi z lahkoto spreminjamo v sivinske ali barvne. S tem pa se spreminja tudi njihova velikost za shranjevanje: barvne slike zavzamejo trikrat več prostora kakor sivinske. Intra- in ekstrakranialne krvavitve so krvavitve v lobanjski votlini in so pogosto posledica travmatske poškodbe. Najbolje jih prikažemo z nativnim CT glave. Pljučna embolija je patologija pljučnih arterij, ki jih zamaši strdek, načeloma z izvorom v spodnjih okončinah. Povzroči zaporo prekrvavitve dela pljuč. Namen: V raziskavi smo ugotavljali, ali se patologija na CT-slikah bolje zazna v barvnih paletah kot v sivinah. Metode dela: Iz baze slik na Zdravstveni fakulteti smo izbrali 40 različnih CT-preiskav: 20 CTA PA in 20 CT glave. V vsaki skupini smo izbrali 10 patoloških in 10 zdravih pacientov. Sivinske slike smo v syngo.vii z LUT-tabelami pretvorili v zeleno in rdečo. Naredili smo anonimiziran in randomiziran seznam 120 slik za vsakega ocenjevalca posebej. V raziskavi je sodelovalo 5 ocenjevalcev, od tega 4 radiološki inženirji in en specialist radiologije. V samem eksperimentu smo jih spraševali o prisotnosti patologije, primernosti prikaza anatomskih struktur in v kateri barvni paleti se jim zdi, da patologijo najbolje vidijo. Rezultati: Ugotovili smo, da barvna paleta ne vpliva pomembno na vidljivost patologije, čeprav smo dokazali statistično značilne razlike v vidljivosti pljučne embolije med nekaterimi barvnimi paletami. Razprava in zaključek: Že narejene preiskave so potrdile, da barve pomembno ne vplivajo na zaznavanje patologije, čeprav bi to teoretično lahko držalo zaradi načina zaznavanja človeškega očesa. Ocenjevalci so najbolj izpostavili zeleno barvno paleto, ki se jim zdi smiselna, da bi jo poleg sive uporabljali v praksi.

• Language: Slovenian
• Keywords: diplomska dela, radiološka tehnologija, računalniška tomografija (CT), barvne palete, zaznavanje barv, zaznavanje patologije
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: ZF - Faculty of Health Sciences
• Place of publishing: Ljubljana
• Publisher: [P. Kelc Hočevar : R. Kožuh]
• Year: 2023
• Number of pages: 52 str., [1] str. pril.
• PID: 20.500.12556/RUL-147312
• UDC: 616-07
• COBISS.SI-ID: 157574403
• Publication date in RUL: 30.06.2023

Secondary language

• Language: English
• Title: Perception comparison of pathologies on CT images of head and thorax in different colour palettes : diploma work

Abstract

Introduction: The human eye perceives many more shades of colour (about a million) than shades of gray (about 30). Different colours affect the individual differently due to emotional and cognitive associations. Images can be easily converted to grayscale or colour using various algorithms. This also changes their storage size: colour images take up three times more space than grayscale images. Intra- and extracranial hemorrhages are bleedings in the cranial cavity and are often the result of traumatic injury. They are best shown with a native CT scan of the head. Pulmonary embolism is a pathology of the pulmonary arteries blocked by a clot, generally originating in the lower extremities. It causes blockage of the blood supply to part of the lungs. Purpose: In the research, we determined whether pathology on CT images is better detected in colour palettes than in grayscale. Methods: We selected 40 different CT examinations from the image database at the Faculty of Health Sciences: 20 CTA PA and 20 CT head. We selected 10 pathological and 10 healthy patients in each group. Grayscale images were converted to green and red in syngo.via with LUT tables. We made an anonymized and randomized list of 120 images for each rater separately. 5 evaluators participated in the research, including 4 radiological engineers and one radiology specialist. In the experiment itself, we asked them about the presence of pathology, the appropriateness of displaying anatomical structures, and in which colour palette they think they can see the pathology best. Results: We discovered that the colour palette does not significantly affect the visibility of the pathology, although we demonstrated statistically significant differences between specific colour palettes in images of pulmonary embolisms. Discussion and conclusion: The investigations that have already been carried out have confirmed that colours do not significantly affect the perception of pathology, although this could theoretically be true due to the way the human eye perceives shades. The evaluators highlighted the green colour palette the most. They think it would make sense for it to be put into practise alongside the existing gray palette.

• Keywords: diploma theses, radiologic technology, computed tomography (CT), colour palettes, colour perception, pathology detection