Key Frame Extraction using Uniform Local Binary Pattern

Key frames are useful in a wide variety of applications like summarizing, storing, indexing and retrieving video clips etc. In this paper, we present an efficient key frame extraction technique. The present method detects key frames efficiently by extracting Uniform Local Binary Pattern of frames of the video. The distance between uniform local binary pattern of consecutive frames is calculated and compared with threshold to extract key frames. We tested the proposed method on categories of serials and news videos. Performance evaluation is carried out using the parameters-figure of merit, detection percentage, accuracy and missing factor. The proposed method offers greater detection percentage. The furnished comparative results analysis proven the efficiency of the proposed method over the state-of-art-methods.