Recent years have witnessed a dramatical growth of the deployment of vision-based surveillance in public spaces. Automatic summarization of surveillance videos (ASOSV) is hence becoming more and more desirable in many real-world applications. For this purpose, a novel frame-selection framework is proposed in the present paper, which has three properties: 1) un-supervision: it can work without requirements of any supervised learning or training; 2) efficiency: it can work very fast, with experiments demonstrating efficiency faster than real-timeness and 3) scalability: it can achieve a hierarchical analysis/overview of video content. The performance of proposed framework is systematically evaluated and compared with various state-of-the-art frame selection techniques on some collected video sequences and publicly-available ViSOR dataset. The experimental results demonstrate promising performance and good applicability for real-world problems.

Unsupervised, efficient and scalable key-frame selection for automatic summarization of surveillance videos

Due to the availability of increased Internet bandwidth and transfer media, the multimedia objects usage is widely used. In multimedia the text files, scripts, images and videos, audios are used by almost every end of users. Daily around millions of videos are being uploaded through various sources. It is becoming very tedious to manage with these multimedia objects specially related to videos. It needs some kind of retrieval technique which will relieve users from tracking videos relatively similar manually. So users are shifted from text based video retrieval to content based video retrieval (CBVR). CBVR has steps as key frame extraction, feature vector formation, similarity and template matching and finally get the retrieved relatively correct expected copies of videos approximately matching with query video. Key frame extraction play very vital role in whole process of CBVR as it decides the search region for search engines affecting the performance. In videos large number of frames forms a scene. This scene includes repetition of nearly same frames with slight differences, which increases the storage space and decreases the performance in video processing. Here instead of searching the whole set of frames in videos, only selected key frame are used for further processing. Automated extraction of key frames has recently gained momentum in the field of video content summarization. Key frames are extracted by computing the consecutive frame differences. Then their relative neighboring frames which nearly match with the key frames are taken aside as near duplicate copies. Therefore the near duplicate copies detection may help to improve the performance and speed of near duplicate detection in content based video retrieval.

An optimized key frame extraction for detection of near duplicates in content based video retrieval

Due to advance growth in videos available across the internet, it is required to navigate and handle them properly. It is essential to select only valuable and accurate information from video. Video summarization helps in acquiring essential information. Video summary produces concise and exact data of the video. With help of key frame extraction video summary can be generated. Key frames from video represent main content of video. In the proposed methodology, to extract key frame from video, haar wavelet transform with various levels and Thepade's sorted pentnary block truncation coding is used. For experimentation purpose test bed of 30 videos is used here. To measure the diversity among successive frames various similarity measures are used. Alias Canberra distance, Sorencen distance, Wavehedge distance, Euclidean distance and mean square error similarity measures are used. The Euclidean distance has given better performance. The increase in accuracy is observed till Haar wavelet of level 5, then higher levels have shown drop in accuracy.

Summarization with key frame extraction using thepade's sorted n-ary block truncation coding applied on haar wavelet of video frame

In the current era, most of the digital information is in the form of multimedia with a giant share of videos. Videos do have audio and visual content where the visual content has number of frames put in a sequence. Most of the consecutive frames do have very little discriminative contents. In video summarization process, several frames containing similar information do need to get processed. This leads to redundant slow processing speed and complexity, time consumption. Video summarization using key frames can ease the speed up of video processing. In this paper, novel key frames extraction method is proposed with Kekere's Proportionate Error (KPE) codebook generation techniques of vector quantization with ten different codebook sizes and two color spaces (RGB and KLUV). Experimentation done with help of the test bed of videos has shown that higher codebook sizes of KPE have given better completeness in key frame extraction for video summarization. The LUV color space with Euclidean Distance with 512 codebook size gives best performance. In square chord Distance, Mean Square Error and Euclidean Distance LUV color space gives better completeness than RGB color space for proposed KPE based video Key frame Extraction

Novel video keyframe extraction using KPE vector quantization with assorted similarity measures in RGB and LUV color spaces

A video is made up of frames. Generally few video processing applications demand to process each video frame one by one, but processing each frame consumes lot of video content summarization helps in improvising the processing speed for such applications. Key frames in video are considered for content summarization. Key frame is a frame in which there is a major change as compared to the previous video frames. Hence key frame extraction becomes very important in Video Content Summarization. In applications needing content summarization, like data storage, retrieval and surveillance, key frames extraction plays a vital role. The Block truncation Coding is one of color feature extraction methods in Content Based Video Retrieval (CBVR). The extended version of BTC is Thepade's Sorted Ternary BTC (TSTBTC). This paper explains key frame extraction using Assorted similarity measure and TSTBTC.

Novel visual content summarization in videos using keyframe extraction with Thepade's Sorted Ternary Block truncation Coding and Assorted similarity measures

Shuping Yang

Papers

Key frame extraction using unsupervised clustering based on a statistical model