Tipizacija vrste tkiva u kroničnim ranama na temelju digitalne fotografije rane

doctoral thesis

• Author: Dubravko Huljev; MBZ: 201863
• Mentor: Tomislav Anić (mentor)
• Mentor: Davor Antonić (komentor)
• Committee member: nije dostupno (član povjerenstva)
• Granter: University of Zagreb; School of Medicine; Zagreb
• Defense date and country: 2011-04-11, Croatia
• Scientific / art field, discipline and subdiscipline: BIOMEDICINE AND HEALTHCARE; Clinical Medical Sciences
• Universal decimal classification (UDC): 61 - Medical sciences

Abstract

Liječenje bolesnika s kroničnim ranama zahtjeva izuzetno kompleksni pristup koji zavisi o više čimbenika: vrsti tkiva u rani (nekrotično, fibrinsko i granulacijsko tkivo), kao i njihovoj zastupljenosti, količini i vrsti eksudata, bakterijskom statusu rane itd. Odnosi između vrste tkiva u rani, uz površinu rane, važan su determinirajući čimbenik u procjeni cijeljenja rana i odluci o terapiji. Precizno mjerenje omogućava efikasniju kliničku procjenu i evaluaciju uspješnosti tretmana. Sa znanstvenog aspekta potrebno je sakupiti podatke kojima se mogu objektivizirati klinička poboljšanja cijeljenja rane tijekom određenog postupka. Primarni cilj ove doktorske disertacije je izrada zlatnog standarda za procjenu točnosti računalnog programa za analizu tkiva u rani. Sekundarni ciljevi su definiranje skupa podataka za učenje klasifikatora računalnog programa koji će na temelju fotografije rane izvršiti automatsku analizu vrste tkiva u rani i njihov udio u cjelokupnoj rani. U tom cilju učinjena su mjerenja subjektivne procjene zastupljenosti tkiva na temelju digitalne fotografije rane. Osam ispitivača je procjenjivalo zastupljenost tkiva na 40 fotografija rane u pet navrata. Greške u subjektivnoj procjeni iznosile su do 36 % u procjeni fibrina, i do 28 % u procjeni granulacija. Objektivno mjerenje istih fotografija učinio je jedan iskusan ekspert u tri navrata, te su srednje vrijednosti mjerenja uzete kao objektivne vrijednosti vrste i zastupljenosti tkiva u rani. Objektivizirano mjerenje udjela pojedinih vrsta tkiva pokazalo je najbolju točnost, preciznost i pouzdanost. Varijacija greške je ispod 5% ukupne varijance. Time je stvoren zlatni standard za provjeru i testiranje računalnog programa za automatsku diferencijaciju i procjenu zastupljenosti tkiva u rani na temelju digitalne fotografije rane. Set za testiranje sastoji se od 40 digitalnih fotografija kroničnih rana. Za svaku fotografiju izračunata je zastupljenost svih vrsta tkiva. Očekivana moguća greška iznosi 5 %. Statističku značajnost razlika između referentnih vrijednosti objektivnog - računalnog mjerenja slika i mjerenja računalnim programom WITA testirani t-testom parova pokazuju da niti za jedan tip tkiva nema statistički značajnih razlika između tih mjerenja.

Abstract (english)

The treatment of patients with chronic wounds request exceptionally complex approach which depends on several factors: type of tissue in the wound (necrotic, fibrin and granulation tissue), as well as their representation, quantity and type of exudates, bacterial status of wound, etc. The relationship between tissue in the wound, with wounds surface, are important determining factor in assesment of wounds healing and decisions about therapy. Precisely measurement allows efficient clinical assessment and evaluation of successful treatment. From the scientific aspect it’s important to collect all data which can be objectify clinical improvement of wound healing during the particular procedure. Primary goal of this doctoral dissertations is making the gold standard for estimation accuracy computer software for wound tissue analysis. Secondary goals are defining dataset for learning software classifier which will make automatic analysis of wound tissue and wound tissue represenrtation, based on digital photography of wound. In that order, measurements of subjective assessment represented tissue are done based on digital images of wound. Eight examiners were assessing representation tissue on 40 images of wounds in 5 occasions. Errors in subjective assessment amounted to 36% in assessment fibrins, and to 28 % in assessment granulation. Objective measurements of the same images were done by experienced expert in 3 occasions. Mean measurement values were taken like objective values of tissue type and wound tissue representation. Objective measurements of wound tissue representation show the best accuracy, precision and reliability. Variation error was under 5% of total variance. This created a golden standard for checking and testing of software for automatic differentiation and assessment of the presence of tissue in the wound based on digital images of the wound. Set for the test consists of 40 digital images of chronic wounds. For each image was calculated representation of all types of tissue. Estimated possible error is 5%. Statistic significant difference between the referent value of the objective - computer image measurement and measurement with computer program WITA tested with pairs of t-test showed that no single type of tissue there was no statistically significant difference between these measurements.

• Keywords: medicina
• Keywords (english): medicine
• Language: croatian
• URN:NBN: urn:nbn:hr:105:026737
• Study programme: Title: Biomedicine and Health Sciences; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, područje biomedicine i zdravstva (doktor/doktorica znanosti, područje biomedicine i zdravstva)
• Type of resource: Text
• Extent: 133 str.
• File origin: Born digital
• Access conditions: Open access
• Created on: 2023-05-05 06:36:02