اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدSearching in one billion vectors: re-rank with source coding
52
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Recent indexing techniques inspired by source coding have been shown successful to index billions of high-dimensional vectors in memory. In this paper, we propose an approach that re-ranks the neighbor hypotheses obtained by these compressed-domain indexing methods. In contrast to the usual post-verification scheme, which performs exact distance calculation on the short-list of hypotheses, the estimated distances are refined based on short quantization codes, to avoid reading the full vectors from disk. We have released a new public dataset of one billion 128-dimensional vectors and proposed an experimental setup to evaluate high dimensional indexing algorithms on a realistic scale. Experiments show that our method accurately and efficiently re-ranks the neighbor hypotheses using little memory compared to the full vectors representation.
قيم البحث
اقرأ أيضاً
The large number of user-generated videos uploaded on to the Internet everyday has led to many commercial video search engines, which mainly rely on text metadata for search. However, metadata is often lacking for user-generated videos, thus these vi
deos are unsearchable by current search engines. Therefore, content-based video retrieval (CBVR) tackles this metadata-scarcity problem by directly analyzing the visual and audio streams of each video. CBVR encompasses multiple research topics, including low-level feature design, feature fusion, semantic detector training and video search/reranking. We present novel strategies in these topics to enhance CBVR in both accuracy and speed under different query inputs, including pure textual queries and query by video examples. Our proposed strategies have been incorporated into our submission for the TRECVID 2014 Multimedia Event Detection evaluation, where our system outperformed other submissions in both text queries and video example queries, thus demonstrating the effectiveness of our proposed approaches.
Image retrieval based on deep convolutional features has demonstrated state-of-the-art performance in popular benchmarks. In this paper, we present a unified solution to address deep convolutional feature aggregation and image re-ranking by simulatin
g the dynamics of heat diffusion. A distinctive problem in image retrieval is that repetitive or emph{bursty} features tend to dominate final image representations, resulting in representations less distinguishable. We show that by considering each deep feature as a heat source, our unsupervised aggregation method is able to avoid over-representation of emph{bursty} features. We additionally provide a practical solution for the proposed aggregation method and further show the efficiency of our method in experimental evaluation. Inspired by the aforementioned deep feature aggregation method, we also propose a method to re-rank a number of top ranked images for a given query image by considering the query as the heat source. Finally, we extensively evaluate the proposed approach with pre-trained and fine-tuned deep networks on common public benchmarks and show superior performance compared to previous work.
This paper describes a new set of block source codes well suited for data compression. These codes are defined by sets of productions rules of the form a.l->b, where a in A represents a value from the source alphabet A and l, b are -small- sequences
of bits. These codes naturally encompass other Variable Length Codes (VLCs) such as Huffman codes. It is shown that these codes may have a similar or even a shorter mean description length than Huffman codes for the same encoding and decoding complexity. A first code design method allowing to preserve the lexicographic order in the bit domain is described. The corresponding codes have the same mean description length (mdl) as Huffman codes from which they are constructed. Therefore, they outperform from a compression point of view the Hu-Tucker codes designed to offer the lexicographic property in the bit domain. A second construction method allows to obtain codes such that the marginal bit probability converges to 0.5 as the sequence length increases and this is achieved even if the probability distribution function is not known by the encoder.
We determine the rate region of the vector Gaussian one-helper source-coding problem under a covariance matrix distortion constraint. The rate region is achieved by a simple scheme that separates the lossy vector quantization from the lossless spatia
l compression. The converse is established by extending and combining three analysis techniques that have been employed in the past to obtain partial results for the problem.
Analog coding decouples the tasks of protecting against erasures and noise. For erasure correction, it creates an analog redundancy by means of band-limited discrete Fourier transform (DFT) interpolation, or more generally, by an over-complete expans
ion based on a frame. We examine the analog coding paradigm for the dual setup of a source with erasure side-information (SI) at the encoder. The excess rate of analog coding above the rate-distortion function (RDF) is associated with the energy of the inverse of submatrices of the frame, where each submatrix corresponds to a possible erasure pattern. We give a partial theoretical as well as numerical evidence that a variety of structured frames, in particular DFT frames with difference-set spectrum and more general equiangular tight frames (ETFs), with a common MANOVA limiting spectrum, minimize the excess rate over all possible frames. However, they do not achieve the RDF even in the limit as the dimension goes to infinity.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
