In the field of content based videoretrieval, recent work has focused on creation and matching of signatures as an effective method for video search, copy or "near-duplicate" detection. In this paper, we ad...
详细信息
ISBN:
(纸本)9781479983407
In the field of content based videoretrieval, recent work has focused on creation and matching of signatures as an effective method for video search, copy or "near-duplicate" detection. In this paper, we address the question of how much content is enough to initiate a content based videoretrieval query from a mobile device. We extract and utilize the motion vectors from an HEvC-encoded bitstream and use them as the content-dependent features to create video signatures. We propose an energy function to quantify the mount of motion information contained in video. Our experimental results exhibit that our proposed approach is able to generate a signature robust to common signal processing techniques and that our proposed energy function can be efficiently used to select the length of a video clip used to query the video database from a mobile device.
Multimedia data grows fast due to advances in information technologies, creating the demand for efficient video indexing and object retrieval techniques. Traditional methods consume significant computational resources...
详细信息
Multimedia data grows fast due to advances in information technologies, creating the demand for efficient video indexing and object retrieval techniques. Traditional methods consume significant computational resources such as storage space and processing time. In this paper we propose an efficient content-based videoretrieval system that is based on three main stages. The first stage involves computing the DCimage of each I-frame, from which a summarization process to extract key-frames is performed. During the second stage, a segmentation processes is applied to each key-frame in order to isolate the region of interest within it. Local features are extracted from the resulting area and are stored as the descriptor of the frame. The retrieval stage is carried out by computing the Euclidean distance and determines if its content is related with the video database. Experimental results show that the proposed approach is promising in terms of efficiency and effectiveness.
Due to the increasing variety and quantity of data in databases, retrieving the desired images among massive images storage becomes a challenge. Hence, many imageretrieval methods are applied on one or some static da...
详细信息
ISBN:
(纸本)9781467365079
Due to the increasing variety and quantity of data in databases, retrieving the desired images among massive images storage becomes a challenge. Hence, many imageretrieval methods are applied on one or some static datasets and the steps of features extraction and similarity comparison are performed on the dataset images as offline. To address the challenge, we propose an online content-based imageretrieval (CBIR) system from huge datasets by applying MapReduce distributed computing model. In the proposed method, images features and their similarity comparison are computed during the retrieval stage. In feature extraction step, similar to most large-scale imageretrieval systems, we employ the bag-of-words model to extract the color and edge histograms from images. Experimental results on the Corel dataset demonstrate that the proposed method improves retrieval accuracy in comparison to the state-of-the-art methods significantly and it is flexible against each database.
Towards effective and efficient image matching or retrieval tasks, the emerging MPEG standard, named Compact Descriptors for visual Search (CDvS), has fulfilled compact descriptors for still images, consisting of comp...
详细信息
ISBN:
(纸本)9781479983407
Towards effective and efficient image matching or retrieval tasks, the emerging MPEG standard, named Compact Descriptors for visual Search (CDvS), has fulfilled compact descriptors for still images, consisting of compressed local and global descriptor. Nevertheless, the frame-level coding of CDvS descriptors from a video sequence does not address the inter-frame redundancy issue, which may consume considerable bandwidth and storage resources. In this work, we propose an efficient coding framework of CDvS descriptors to generate compact descriptors for video sequences. For local descriptors, we propose a multiple reference predictive technique to exploit the temporal correlation of local descriptors and location coordinates over a sequence of frames. To further improve the prediction performance, keypoint tracking is applied to identify temporally repeated keypoints. For global descriptors, a propagation coding way is employed to compress the global descriptors of adjacent frames. The empirical evaluation has shown that the proposed coding approach has yielded a low bit rate of less than 40kbps on average, while maintaining comparable matching and retrieval performance. Compared to the sequence of original frame-level CDvS descriptors, the proposed approach has achieved over 25× bit rate reduction.
Human pose as a query modality is an alternative and rich experience for image and videoretrieval. We present a novel approach for the task of human pose retrieval, and make the following contributions: first, we int...
详细信息
Human pose as a query modality is an alternative and rich experience for image and videoretrieval. We present a novel approach for the task of human pose retrieval, and make the following contributions: first, we introduce `deep poselets' for pose-sensitive detection of various body parts, that are built on convolutional neural network (CNN) features. These deep poselets significantly outperform previous instantiations of Berkeley poselets [2]. Second, using these detector responses, we construct a pose representation that is suitable for pose search, and show that pose retrieval performance exceeds previous methods by a factor of two. The compared methods include Bag of visual words [24], Berkeley poselets [2] and Human pose estimation algorithms [28]. All the methods are quantitatively evaluated on a large dataset of images built from a number of standard benchmarks together with frames from Hollywood movies.
Cloud computing is collection of distributed hosts which allow services on demand to user. The service that provide to the user will include multimedia such as image/videoretrieval. More number of users demands vario...
Cloud computing is collection of distributed hosts which allow services on demand to user. The service that provide to the user will include multimedia such as image/videoretrieval. More number of users demands various multimedia computing resources and storage services through internet at same time in order to resolve this cloud based multimedia system emerged. The CMS will receives client's request for multimedia service task through resource manager and assign those task requests to different server clusters according to the characteristics of the requested tasks. In the existing system genetic algorithm is used for resource allocation where the videos get spitted and they are uploaded to server where the resource manager will retrieve the video files which has the drawbacks as it consumes more time and increase complexity. In the proposed system Fibonacci based splitting strategy used to split a video file into number of chunks that allows to reduce the delay, optimize the resource utilization and increases the performance by fast video download.
Large amounts of databases are created due to the developments in data storage and acquisition technologies. There is a need to develop an appropriate system that will manage these entire databases. Also we need to pr...
详细信息
ISBN:
(纸本)9781479953646
Large amounts of databases are created due to the developments in data storage and acquisition technologies. There is a need to develop an appropriate system that will manage these entire databases. Also we need to precisely and effectively retrieve images from these databases for various applications. The Content Based imageretrieval (CBIR) system serves this purpose. In this paper, we introduce a user based system for CBIR in which genetic algorithm is applied. The different features of color image such as mean, standard deviation and the image bitmap are used for retrieval. In addition, the texture features such as the edge histogram of an image and the entropy of the gray level co-occurrence matrix are used. Furthermore, the genetic algorithm is applied to help the user in identifying the images which satisfy his needs for reducing the gap between the users' expectation and the retrieval results.
We study the challenges of image-based retrieval when the database consists of videos. This variation of visual search is important for a broad range of applications that require indexing videodatabases based on thei...
详细信息
Many of the Internet applications such as video conferencing, military imagedatabases, personal online photograph albums and cable television require a fast and efficient way of encrypting images for storage and tran...
详细信息
Many of the Internet applications such as video conferencing, military imagedatabases, personal online photograph albums and cable television require a fast and efficient way of encrypting images for storage and transmission. In this paper, discrete logarithms are used for generation of random keys and Number Theoretic Transform (NTT) is used as a transformation technique prior to encryption. The implementation of NTT is simple as it uses arithmetic for real sequences. Encryption and decryption involves the simple and reversible XOR operation of image pixels with the random keys based on discrete logarithms generated independently at the transmitter and receiver. Experimental results with the standard bench mark test images proposed in the USC-SIPI data base confirm the enhanced key sensitivity and strong resistivity of the algorithm against brute force attack and statistical crypt analysis. The computational complexity of the algorithm in terms of number of operations and number of rounds is very small in comparison with the other image encryption algorithms. The randomness of the keys generated has been tested and is found in accordance with the statistical test suite for security requirements of cryptographic modules as recommended by National Institute of Standards and Technology (NIST).
Multimedia database can be define as a collection of storage and retrieval systems, in which large amount of media objects are created, modified, searched and retrieved, where as Multimedia is the combination of text,...
详细信息
ISBN:
(纸本)9781479928996
Multimedia database can be define as a collection of storage and retrieval systems, in which large amount of media objects are created, modified, searched and retrieved, where as Multimedia is the combination of text, image, graphics, animations, audio and video. The extension of database application to handle multimedia objects requires synchronization of multiple media data streams. Multimedia data mining refers to the extraction of implicit knowledge, data relationships, or other patterns which are not stored in multimedia files explicitly. The system's overall performance in retrieval can be increase by indexing and classification of multimedia data with efficient information fusion of the different modalities is mandatory. Apart from text retrieval, the current waves in web searching and multimedia data retrieval are the search for and delivery of 3D scenes, images, music and video. The content-based multimedia information retrieval provides new techniques and methods for searching various multimedia databases over the world.
暂无评论