7.1 Introduction

Video is a visual media source that joins a succession of pictures to frame. The video transmits a frame and the procedures based on the request in which the frame catches should be appeared. Sound segments are compared with the frame being appeared on the screen. Streaming means sending either audio or video data, which enables to begin its processing before it is totally received. Video streaming is a sort of media streaming in which the information from a video document is constantly conveyed by means of the Internet to a remote client [2]. It permits a video streaming to be seen online without being downloaded on a host PC or gadget. Video streaming takes a shot at information streaming standards, where all video streaming recorded information is compacted and sent to a gadget in little lumps. Video streaming normally requires a good video player that connects with a remote server, which has a pre-recorded media document or live feed [10]. The server uses specific calculations to pack the media record or data for exchange over the system or Internet federation.

7.2 Existing Technology and Its Review

In past decades, some ongoing video streaming conventions and mechanisms have actualized like the Real Time Streaming Protocol (RTSP) and Real Time Messaging Protocol (RTMP). These protocols cause it workable to the clients to distribute sound and video streams up the network and perform them frequently. The RTSP is an organized governing protocol proposed to use in amusement and correspondence frameworks to organize streaming media servers. This convention is utilized for implementing and governing media sessions across endpoints. It has been actualized in QuickTime Streaming Server like Apple’s shut source streaming server. Continuous Messaging Protocol (RTMP) is a protocol created by Macromedia (now claimed by Adobe) for streaming sound, video, and information over the internet. This convention is utilized for communication between a Flash Player and Adobe Flash Media Server. RTMP sessions might be scrambled utilizing SSL or using RTMPE, however do not give adequate security. Adobe Flash Player can be downloaded for nothing on the client side. Adobe likewise gives engineer rendition of Flash Media Server for non-business purposes for secure streaming which utilizes SSS (Secure Scalable Streaming) which fragments the video outlines into tiles and afterward codes the tiles into header and versatile information arrange, then packetize the header and encode information.

For continuous media streaming on cell phones, RTMP has a couple of problems, RTMP is a TCP-based convention, which underpins retransmission for lossless data correspondence. This implies framework architects need to combine a good missing information retransmission procedure to dodge sound and video delay, jitter, and asynchronization on account of the information parcel misfortune. Besides, RTMP uses distinctive conventions/ports from HTTP that causes helpless in receiving hindered by firewalls and works with Flash.

RTSP is a content based application-layer convention. It assumes the job of “network remote control” in sight and sound administrations, for example, sound and video progressively. In the media streaming transmission, RTSP includes a couple of essential protocols for information communication, media control, and media size depiction. Most RTSP servers exploit Real-time Transport Protocol (RTP) as a distributing technique to broadcast the information streamed by developing a Transmission Control Protocol (TCP) or User Datagram Protocol (UDP) association as a media streaming conveyance channel [4]. RTSP portrays the advancement of guidance for communication of clients and streaming servers containing OPTIONS, DESCRIBE, ANNOUNCE SETUP, PAUSE, PLAY, RECORD, REDIRECT, and TEAR-DOWN. At the circumstance when a client commences a media streaming session to a streaming server through RTSP directions, they exploit the Session Description Protocol (SDP) to trade multimedia subtle elements, ship addresses and other session confession metadata.

TCP is a federation arranged and strong byte-stream tradition. The sender and beneficiary should fabricate a TCP relationship before data transmission. TCP is a vehicle layer protocol. It has course of action instruments to ensure the transmission trustworthiness. Three-way handshake, moreover insinuated as “SYN-SYN-ACK”, is required before transmitting data. In the midst of data transmission, every datum segment should be perceived by the beneficiary. If the sender does not get the perceived information from the authority in a predestined time, TCP will use a retransmission timeout part to qualify non-adversity data an utilize course of action assertion numbers to guarantee the data divide is in the right demand. With an explicit true objective to control the surge of data and to upgrade transmission proficiency, TCP has a sliding window part and uses moderate start calculation to keep up a key separation from organize stop up.

UDP is a connectionless and flawed protocol. Like TCP, UDP is in like manner a vehicle layer tradition. UDP is only in charge of sending and tolerating the datagram, anyway it does not guarantee the datagram is gotten by the objective in the wake of sending the datagram. As needs be, the data can be gotten out of demand, or even lost. Diverged from TCP, UDP is speedier on the grounds that UDP has no stream control, no screw up checking and no datagram recognized systems. Thus, UDP is much of the time used by sight and sound applications for transmitting the data stream, for instance, sound and video streaming data in light of the fact that these applications are asking for on nonstop response and can’t be conceded.

RTP is an internet transport protocol which handles real-time multimedia information streams communication. It is determined as functioning in the transport layer, constructed above UDP. The use of RTP is to provide time information and to synchronize several streams. RTP only assures real-time data communication but does not assist a trustable transport mechanism for transmitting data packets in a sequence [4]. Also, it does not give flow control and traffic control, which depends on Real-Time Transport Control Protocol (RTCP). RTP gives a timestamp, serial number, and other structures to handle the real-time streaming data. After getting data packets, the client retrieves the data packets in the actual sequence as per the RTP header information that conveys the customers how to retrieve the data packets and how the codec bit streams are unloaded. RTP header information contains timing information, sequence number, payload recognition, frame indication, source recognition, intra media synchronization, etc.

RTCP is a control protocol which is subject for handling transmission standard between applications to interchange control details on vast networks, mostly for streaming media, phone and video conferencing. In the course of an RTP session, an application utilizes two ports such as RTP and RTCP, respectively. RTCP packets are sent systematically to monitor the standard of the service and change of user’s session details and other functions. RTCP packet comprises the number of packets sent, lost and other data via receiver report (RR), sender report (SR), source description items (SDES), shows the end of participation (BYE) and application specific functions (APP) packets formats. Therefore, the server can utilize these details to dynamically modify the transmission rate, and also modify the payload type. RTP and RTCP work together to reduce transmission above and improve efficiency.

SDP is used to illustrate multimedia sessions. It serves for session statement, request and other forms. SDP does not help the settlement operation of SDP session information or media encoding. When commencing audio/ video streaming, video call, or other sessions, there is a need to deliver media information, transport addresses and other session information metadata to the participants. SDP gives a standard portrayal for such details such as session name and objectives, session time, media time and more.

Extensible Messaging and Presence Protocol (XMPP), also known as Jabber convention was planned for IM and online presence detection. XMPP depends on Extensible Markup Language (XML) streaming technology. XMPP makes informing over the web possible, independent of working frameworks and programs. XMPP is intended to support IM undertakings, for example, confirmation, get to control, end-to-end encryption and compatibility with different conventions. Moreover, the XMPP Standards Foundation (XSF) creates numerous ex-strains (XEPs) which make XMPP all the more intense such as roster, client and server elements or traits whose broadened namespaces are “jabber:iq:roster”, “jabber:client”, and “jabber:-server”, separately. Due to the fast evolution of wireless networks and smart technologies, mobile video streaming and social networks have become an important part of the lives in various areas (social connections, education, entertainment, scrutiny, etc.). There have been many video streaming technologies on mobile devices over the web, from the client to streaming server side, from video compression to streaming protocols.

Most of the studies and applications made on video streaming are concentrated on video-clip distribution and currently popular videos streaming. Some are concentrated on video-clip or real-time notations as available in video sharing. Cheng and Liu submitted a NetTube, for small video sharing application. They created a model to minimize the server workload to optimize the playback standard and scalability.

Jia and Ma submitted MoviShare, a video distributing stage that can give the mobile users with video searching and posting services. MoviShare targets a logical combination of area-based mobile social connecting and multimedia distribution. It can likewise create a GIF (Graphics Interchange Format) record of accessible video cuts as video abstraction to take care of transmission capacity and power restriction issues [5]. However, MoviShare is certainly not a live streaming framework yet an on-demand streaming. Clients can only share their videos once they are recorded and the same is sent to a server.

Silva et al. portrayed a technique for explaining objects on a live video streaming on Tablets. They planned two ways to deal with enable clients to include comments when the object is moving in the video streaming. They utilize object tracking techniques, for example, Kinect sensor to make stays on those explanations to maintain a strategic distance from the comments lost when overlaid. Yamamoto et al. proposed an approach to create explanations dependent on social movements in association with video cuts, for example, client remarks and weblog.

They built up a framework called Synvie to separate profitable data from those social and network exercises as video streaming explanations. El-Saban et al. displayed a framework for constant video explanation of captured recordings on cell phones to encourage perusing and seeking. A client can utilize this framework to capture a video using cell phone. The video is sent, continuously, to a brought together server which breaks down video key casings to create explanations by utilizing MSERs (Maximally Stable External Regions) detector with SIFT (Scale-Invariant Feature Transform) highlights. At that point these explanations are returned back to the client’s cell phone. Their work is centered on producing explanations from constant recordings.

H. Sun et al. displayed a review of video streaming systems which incorporate scalable video coding, video transcoding, network protocols, and streaming techniques that have been produced as of late. Walker et al. gave a design of mobile video streaming under 3G and GPRS systems and some standard methods for sound/video streaming, for example, sound/video compression and streaming protocols. In their framework, they utilized customer buffering for packet loss and stream switch-down choice for lower bit rate. In any case, their framework has a consistent postponement of many seconds when streaming live occasions. They tried it with just a single client.

Meyer tended to an examination take a shot at video streaming with Client/Server design. He broadened the SpyDroid 2 project and utilized the VLC Player 3 as a customer streaming player, although his center is advancing encoding principles as per the nature of association for transmission and limiting utilization of energy. It has no social exercises or touch display communications.

7.3 Methodology and Research Design

Video is a visual multimedia basis that associates a series of pictures to form a motion picture. The video spreads a signal to a screen and processes the sequence in which the screen captures should be displayed. Videos usually have audio elements that match with the pictures being presented on the screen.

Video Retrieval is wide area that incorporates features from numerous features and fields including artificial intelligence, machine learning, data base management systems, etc. There have been huge numbers of algorithms fixed in these fields to achieve various video retrieval responsibilities. In our Proposed System video retrieval is nearly a subdivision of Internet Multimedia Subsystems (IMS).

Frames are individual pictures in a sequence of images (video). A key-frame or a unique frame is a frame used to indicate the beginning or end of a change made to a parameter. An arrangement of keyframes describes which movement the spectator will see, while the location of the keyframes on the video expresses the timing of the movement.

Feature extraction normally refers to the process of extracting features from a frame in a video, independently of past or future frames. Feature extraction includes decreasing the amount of resources needed to define a huge set of data.

Feature extraction is a significant part of the system. From the key frame, feature extraction is done. The features of the key frames are extracted according to the requirement from the video. Classification is performed using classifier from the extracted features. In this system, Support Vector Machine (SVM) classifier can be used. SVMs are binary classifiers, yet, they can be accepted to handle the various classification tasks common in remote sensing. SVM is used in information retrieval. The extracted features of the keyframes undergo SVM, where, the classifier classifies various features like behavior, color, shape, etc., of the required data.

This system is preliminarily to detect a person who once had been identified as a victim in any previous cases. For this the video footage of the incidental scenario is partitioned into frames, and from all the frames of the footage the one containing a perfect shot of the complainant is identified. The identified frame is used as a key frame and any frame matching with key frame could be used to determine the complainant in future. The environment returns back to normal situation in a short period of time. Since there exists live video streaming, whenever the victim enters again into the environment where the incident have occurred, he/she could be found by matching the current frame with the key frame and can be intimated to the respective people belonging to that environment. A real-time application and prototype for the system is discussed below:

Consider a robbery has happened in bank. From the video footage during the incident a shot of the victim can be identified. The identified frame can be set as the key frame. The bank returns back to a normal situation within few days. Whenever the victim enters the bank after the recovery of bank, he could be identified with the help of key frames found in prior. When the frame of a live video matches with the key frame, the bank employees can be alerted that the identified victim has entered the bank which helps to capture easily. HLS protocol is used to identify the victim.

7.4 Findings

The transmission delay of various protocols are shown in Figure 7.1. The TCP transmission delay is between 500 ms and 620 ms. The RTP transmission delay is between 20 ms and 70 ms. The HLS transmission delay is between 10 ms and 30 ms. The RTMP transmission delay is between 60 ms and 80 ms. The UDP transmission delay is between 50ms and 65ms. Therefore, using HLS protocol to transmit can effectively reduce the transmission delay, and improve the streaming efficiency as shown in Table 7.1.

7.5 Future Research and Conclusion

The crime rate happening in day to day life and with secured protocol of HLS to overcome with such event is discussed. The adopted system is an implementation of video classification method with HLS security protocol, discussion in the previous sections provides encouraging results and comparatively higher efficiency is achieved by using multiple frames instead of single frame or key frames representing a shot. HTTP based live streaming has the better traversal of NAT and firewalls, ease of deployment, and built-in friendly bandwidth sharing which is used for transfer of videos easily with high transmission bit rate remotely to other places [11]. This is a study for identification of crime and criminals and in the society and also for providing a better future to live in.

References