Development of a software system of video-integrated analysis made for the motion detection and analysis of sleep. A

Introduction The study aims to validate the system for the automatic analysis of the video-polygraphic, comparing with traditional methods of visual analysis. In particular, we will: detect motor activity and compare the analysis of PLM by classical methods: video-actigraphy versus Immobilization Test. Analyze the results of motion detection by comparing epoch by epoch. Materials and methods 24 Patients studied with Immobilization Test in sleep laboratory and video analysis. Movement study with video analysis offline through software system composed by ManyCam, WebCamXP, ZoneMinder, ZoneMinderAnalyzer (ZMA), Actiwatch Activity & Sleep Analysis 7. The motion was classified according to the AAMS criteria and calculated the leg movement and the motion sequence. The video analysis was performed with combined use of ManyCam and WebCamXP which permits obtain an IP virtual Camera. ZMA use the provided data by video analysis performed with ZoneMinder. ZM reveals the differences of video frame in input, quantifies them and then stores them in a database SQL. ZMA divides the time in epochs of x seconds and calculates for those epochs compatible values with these obtained through polygraphic analysis used the database data. The obtained data are transformed in a format readable by Psg software analysis. The results obtained were compared with the analysis performed by the system software (Zoneminder). The comparison was made epoch by epoch. Results The video analysis offline offers an approach to study the motion during sleep, not always visible with the classic video system and with EMG electrodes on the surface. This system permits an analysis of body movements, along witih those of the face and the limbs. The ManyCam and WebCamXP system is more complete in comparison on macrostructure sleep, visual analysis of body movements, spectral analysis of EMG and actigraphy analysis. We analyzed a total of 4512 times. Time-period analysis, the coefficient K Cohen has established a degree of agreement between ZM and Immobilization Test equal to 0.252. The Bland–Altman analysis confirmed that scores of ZM were not significantly different from those obtained with test Immobilization. The results of the analysis of Bland–Altman show a substantial overlap between the scores obtained with ZM and Immobilization Test with a slight tendency to overestimate ZM motor events. The Pearson coefficient, equal to 0.2523 (corresponding to a significance value of p Conclusion Our study makes it possible to collect the following conclusions:The motor pattern analyzed with different methods shows a substantial overlap; Zoneminder allows detection of motor events also minimal compared with the analysis carried out visually test of immobilization. By adjusting the sensitivity of motion detection you can ’analyze a specific body area or a specific region. From our study shows an easy applicability of the method especially in long recordings, allowing effective results with cost containment. This method also allows the non-invasive nature of the system on patient therefore making themselves available for recordings of newborns and infants.