We encountered lots of technical problems and we could not handle and solve these problems.

Background and rationale: Ultrasound-guided regional anesthesia is a widely used pain control method today. A critical aspect of the procedure is accurate visualization of anatomical structures on ultrasound to precisely define target areas. Distinguishing surrounding tissues with an imaging model that automatically recognizes sonoanatomy in ultrasound images will reduce unintended intraneural injections or injury to other anatomical structures in close proximity and increase patient safety.

Research question; How can we improve the ultrasound images we frequently use in regional blocks by integrating them with artificial intelligence to reduce complications and improve applications? And what is the accuracy of the developed artificial intelligence support during imaging?

Research purpose; This work; We aim to further increase the safety of different regional block positions, minimize the risk of complications, and improve ultrasound visualization by developing an artificial intelligence model (AI Model-Artificial Intelligence) that automatically identifies and segments anatomical landmarks, provides visual guidance for inexperienced colleagues, and improves the performance of the developed model during application. aims to demonstrate its accuracy.

Hypothesis; Numerous studies have shown that the use of ultrasound and neurostimulators in practice increases the success, onset and quality of nerve blocks, but due to the low incidence of major complications and the absence of comparable randomized studies, no definitive statement can be made as to whether ultrasound reduces the overall rate of nerve damage. An imaging model that automatically marks sonoanatomy with artificial intelligence in ultrasound images can reduce unintended intraneural injections or injury to other anatomical structures in close proximity and improve patient safety.

Research processes; The study will consist of two stages. Phase 1: PENG, collection of sonoanatomical images of Supraingiuinal Fascia Iliaca block and development of artificial intelligence software (in healthy volunteers without invasive procedures) Stage 2: Conducting validation study with the developed artificial intelligence

Stage 1: Taking ultrasound images from healthy volunteers (150 volunteers) to produce artificial intelligence - How to take PENG and Suprainguinal Fascia Iliaca Block sonoanatomical images is as follows.

1.1) PENG (Pericapsular Nerve Group Block): Images will be taken with both linear and convex probes. Sonoanatomical information will be collected from healthy volunteers and no invasive procedures will be performed. 150 (75 women -75 men) healthy volunteers who agree to have ultrasound images taken will be included.

1.2) Suprainguinal Fascia Iliaca Block: Images will be taken with a linear probe. Sonoanatomical information will be collected from healthy volunteers and no invasive procedures will be performed. 150 (75 female-75 male) healthy volunteers who agree to have ultrasound images taken will be included.

In the first phase of this study, thanks to the PENG and Suprainguinal Fascia Iliaca block images collected from volunteers, the artificial intelligence technology Smart Alfa Teknoloji San. recognizes and marks the anatomical structures of this region. and Tic. Inc. It will be developed by and added to Nerveblox software. After PENG and Suprainguinal Fascia Iliaca blocks are included in the software, Nerveblox software will be used during validation in the second phase of the study.

Phase 2: In the second phase of the study, Smart Alfa Teknoloji San. and Tic. Inc. Artificial intelligence technology called Nerveblox, which was developed with the data received in the first stage with the support of the company, will be used. It is the validation and accuracy study of the artificial intelligence technology developed in the first stage. The accuracy study will be conducted on 40 healthy volunteers. 20 men and 20 women will be included in the study.

Thanks to the Nerveblox artificial intelligence technology developed at this stage, the accuracy of the anatomical structures marked and colored by the regional-specific artificial intelligence; It will be evaluated by 6 experienced anesthesiologists based on ultrasound image scans made by 2nd, 3rd and 4th year assistants (two assistants from each year). The second phase is the validation phase and the validators will be experienced anesthesiologists (at least five years of specialized experience). Validators will score the accuracy of representation of each predefined anatomical landmark using a 5-point scale (1: Very Poor, 2: Poor, 3: Good, 4: Very Good, 5: Excellent); Accuracy is defined as expert opinion on how software-generated landmark labels represent true anatomy in raw ultrasonography images.

Inclusion Criteria:

• Agreeing to take ultrasound images
• Healthy - No comorbidities
• Adult individuals between the ages of 18-65

Exclusion Criteria:

• Those who do not accept ultrasound images
• Individuals under 18 years of age
• with comorbidities
• Pregnancy

150 (75 women -75 men) healthy volunteers who agree to have ultrasound images taken will be included.

For the first phase of the study, 150 volunteers, 75 women and 75 men, were determined by the Smart Alfa company, which will produce artificial intelligence.

In the sample size analysis conducted for the second stage, it was determined that there should be 19 participants for each group with an alpha margin of error of 0.05, with a power rate of 80% for comparison of two groups. The effect used for this calculation was calculated as 0.837, and the actual power was calculated as 0.812, based on similar studies. As a result of the analysis, considering the drop out rate, it was planned to recruit 20 participants for each group. (20 women - 20 men)

We deeen to talk to people who is volunteer for the study and company for that.