REAL TIME VIDEO ACTIVITY DETECTION TECHNIQUES IN MACHINE LEARNINGAuthor :Ms. Ashwini M bhugul, research scholar. Sipna COET, Amaravati.Dr. Vijay S Gulhane, Head & Professor, Sipna COET, Amaravati.ABSTRACT :Real time video activity detection techniques in machine learning is anemergence in the field of the technology due to its most varied applications. Real-time videoactivity detection is more advantageous when it comes to the survey of Navigation andinteraction of things socially. In this research work, we will analyze different types of techniquesthat are used for activity detection in machine learning and hold this will create an advantageplatform for the community and we also discuss the performance of these techniques.INTRODUCTION :The activity detection technique using machine learning is not the recent one But manyresearchers are studying and carrying out a series of studies resolving this matter and sometechniques are even related to the health sector surveillance sector and computer-aided works.The real-time video detection will also offer a wider range of warnings and countermeasures thatcan be taken place whenever the danger occurs in the process and recognition of the process willalso give an ability to detect whether the proposed technique is right or wrong. Computer visionis truly a basic and major technique used in real-time video activity detection which is alsoknown as a video detection technique in machine learning and the visual aids provided by thistechnique are related to the information of the actions and the video events carried out in theprocess. Different techniques are used for video activity detection such as optical flow, deeplearning, and hidden Markov models. Many newer models add also in the studies such as singlerecognition, group recognition, etc. The deeper analysis of the process will also generate themood accurate standard of results and the realization of this task will give the higher reliability tothe video surveillance.LITERATURE REVIEW :Many types of automatic recognition help us to get alert and reliable data. The video activitydetection and I think my sis some patterns such as running smoking walking playing and othercomputerized visual formats in the real-time video form using machine learning[1]There aremany research activities and surveys that have been carried out recently which predict that thepictorial structures in the form of the frame are more commonly used techniques in humanactivities. Due to the contraction of the body parts, we can say that the 2D image will becomplicated manner so recent research said that there can be a newer improvement in the imagein-the depth and to relate that to the 3D data[2].
Deep learning is one of the subset of machine learning so you see that field of study in real-timevideo detection we can work on the sequence of the frames and in the recent developments wecan say that the motion energy image MEI, is taken as the most effective methods because of thesimplicity and the higher accuracy in all the environments[3]. The technique of motion energyimage can be made individual formats such as if we are making visual aid from the top wave andthis surveillance is to the ground[4].
Upload your study docs or become a
Course Hero member to access this document
Upload your study docs or become a
Course Hero member to access this document
End of preview. Want to read all 9 pages?
Upload your study docs or become a
Course Hero member to access this document
