Andmebaasi logo

VLI lõputööd

Selle valdkonna püsiv URIhttp://hdl.handle.net/10492/2470

Sirvi

Sirvi VLI lõputööd Märksõna "AI" järgi

Nüüd näidatakse 1 - 2 2
  • Pisipilt
    Kirje
    The clinical decision-making process and the potential impact of the GekkoVet clinical decision support system on diagnostics in veterinary medicine
    (Eesti Maaülikool, 2023) Heino, Olli Jetro Juhani; Karindi, Martin (juhendaja)
    Clinical decision-making is a multidimensional process, and factors common to any kind of human decision-making cause biases. Errors in clinical decision-making are common and significantly impact individuals and society. In veterinary medicine, as in human medicine, an increased amount of clinical data and information has created a need for clinical decision support systems (CDSS). In this study, was evaluated veterinarians’ ability to build differential diagnosis lists in four real-life internal medicine cases. In each case additional information about the patient was provided in three phases. Forty-seven licensed veterinarians completed the survey. In the CASE 4 over 90% found the correct diagnosis. Seventy-seven percent of the respondents answered correctly for CASE 2, while lower scores were in CASE 1 and CASE 3, with 47% and 55% of respondents providing correct diagnoses, respectively. GekkoVet CDSS got all four cases correctly diagnosed in the last phase of the case (respondent average: 2.7). In total, GekkoVet scored 8 out of 12 possible right answers, whereas 6.19 was the average score among respondents. Those respondents who used external help did spend more time (p-value 0.034) to complete the questionnaire. It is notable that using external sources and more time did not significantly improve respondents’ result. Experience as a practicing veterinarian did not influence the average number of diagnoses, the amount of time spent to complete the survey or the number of correct diagnoses at a statistically significant level. Based on this study and literature published previously, CDSSs have the potential to reduce clinicians’ workloads and help them make more diagnoses faster and with greater confidence. Along with potentially increased clinical performance, CDSSs might reduce the burden veterinarians experience from clients and employers. Certain cautions and limitations should be considered when implementing CDSSs for clinical use.
  • Pisipilt
    Kirje
    Comparison of lameness evaluation through video and AI analysis
    (Eesti Maaülikool, 2026) Salonen, Ilona Eeva Katri; Corrêa, Felipe (advisor); Kliinilise veterinaarmeditsiini õppetool
    Equine lameness is one of the most common clinical problems affecting horses. Although subjective gait assessment remains the gold standard for lameness examination, observer bias underscores the need for objective methods to support it. Artificial intelligence (AI)- based applications have emerged as potential tools for use in clinical settings and veterinary education. The aim of this thesis was to evaluate how accurately participants with limited prior experience in lameness assessment and detection would assess forelimb lameness from recorded videos, using the AI program SLEIP as a reference. Additionally, this thesis aimed to see if the participants would improve after a lecture on lameness detection. Eight first-year veterinary students and five other university members assessed twelve videos of horses. They tried to identify the lame limb and classify the lameness severity. They completed two questionnaires: one before the lecture and a second after it. Results showed a significant improvement in recognition of the affected limb, with the average of correct answers increasing from 4,61 to 8,15, but the severity rating ability improved only slightly. These findings indicate that short education supported by video material can improve recognition of lameness laterality, but severity classification may require additional training.