IMAGE RECOGNITION METHODOLOGY IN VIDEO

Computer Science And Engineering

NATIONAL CONFERENCE ON TECHNOLOGICAL ADVANCEMENTS IN ENGINEERING

Image processing is an active research area, changing and improving constantly. In this methodology, a video is given as input and it collects large number of images from the video. Images with similarity in the video are compared. It identifies and recognizes the given image in the video. It partitions each frame of the matching images to recognize image in video. For each frame in video sequence it first detects the given image in the video. It then partitions all the images in to K-different partitions by K-means algorithm. It computes the look-alike images from the video. It captures variations in pose and expressions of the given image in the video. This technique significantly increases the image recognition and hence, it is an efficient way of identifying the similar images in the video. In existing system, it compares two videos to get the given input image. The video is converted into frames. For the sequence of frames, large number of images is compared in the video to get the correct result. Each frame is processed as a thread. It performs clustering, fusion of images in the video to get the faces of persons those are present in both the videos. Video-based recognition has also gained lot of traction. In video-based face recognition, a key challenge is using the extra information available in a video; e.g. face, body, and motion identity cues. In proposed system, the image recognition system is used to capture the images in the video. In the first step the video is scanned. It is used to compare the image in the video. In image recognition system, image is given as the input which is compared with the video. Video is partitioned into many frames. Video frames are processed in parallel manner. The objective of this research reduces the compilation time by processing the frames in parallel and makes it more effect. Similar images are been compared in the video. It performs the fusion of images. A similar image in the video is scanned to get the correct result. Frames are created after each second. Video is to enhance the computation speed to get the output result in minimum time. In this proposed system, we use K-means algorithm to group the frames that are similar to each other. The K-Means algorithm organizes frames into k partitions.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.