International Symposium/Conference on Music Information Retrieval, 2013
A musical medley is a piece of music that is composed of parts of existing pieces. Manually creat... more A musical medley is a piece of music that is composed of parts of existing pieces. Manually creating medley is time consuming because it is not easy to find out proper clips to put in succession and seamlessly connect them. In this work, we propose a framework for creating personalized music medleys from users' music collection. Unlike existing similar works in which only low-level features are used to select candidate clips and locate possible transition points among clips, we take song structures and song phrasing into account during medley creation. Inspired by the musical dice game, we treat the medley generation process as an audio version of musical dice game. That is, once the analysis on the songs of user collection has been done, the system is able to generate various medleys with different probabilities. This flexibility brings us the ability to create medleys according to the user-specified conditions, such as the medley structure or some must-use clips. The preliminary subjective evaluations showed that the proposed system is effective in selecting connectable clips that preserved chord progression structure. Besides, connecting the clips at phrase boundaries acquired more user preference than previous works did.
During medical treatment, personal privacy is involved and must be protected. Healthcare institut... more During medical treatment, personal privacy is involved and must be protected. Healthcare institutions have to keep medical images or health information secret unless they have permission from the data owner to disclose them. Reversible data hiding (RDH) is a technique that embeds metadata into an image and can be recovered without any distortion after the hidden data have been extracted. This work aims to develop a fully reversible two-bit embedding RDH algorithm with a large hiding capacity for medical images. Medical images can be partitioned into regions of interest (ROI) and regions of noninterest (RONI). ROI is informative with semantic meanings essential for clinical applications and diagnosis and cannot tolerate subtle changes. Therefore, we utilize histogram shifting and prediction error to embed metadata into RONI. In addition, our embedding algorithm minimizes the side effect to ROI as much as possible. To verify the effectiveness of the proposed approach, we benchmarked three types of medical images in DICOM format, namely Xray photography (X-ray), computer tomography (CT), and magnetic resonance imaging (MRI). Experimental results show that most of the hidden data have been embedded in RONI, and the performance achieves high capacity and leaves less visible distortion to ROIs.
Since a new discrete rectan lar wave transform (DRWT) has been propcsef there are various applica... more Since a new discrete rectan lar wave transform (DRWT) has been propcsef there are various applications which could be applied, such as real-time DFT implementation and ima e recognition. Although the DRWT possesses tfe property of easy-computation, there 19 no cost-effective algorithm for the implementation of DRWT. Most of the proposed architectures to implement this transform are systolic array, wavefront array, or algorithm based implementations with butterfly-like structures. Since these proposed architertiires have either the modular or high-throughput rate property, but not both; therefore, a same-geometry implementations B proposed in this paper. In this paper, a cost-effective algorithm for implementing DRWT is roposed, called the s a m T m e t r y DRWT. TKis newly proposed dgorit m provides a better procedure not merely in fast computation but also in the consideration of real implementation. In the sarne-geometry DRWT, power of two length is the cdnstraint. There are 21oaN-1 stages in the implementation of the DRWT with length N. Furthermore, logzN stages are identical and therefore only logzN st es demand circuit design in implementation phase.%he delay of each stage is merely one addition and the cells required to construct these stages are simple. Therefore, this algorithm is suitable for VLSI implementation.
The utilization of real-number DFT codes for channel equalization is studied in this letter. As s... more The utilization of real-number DFT codes for channel equalization is studied in this letter. As shown below, through real-number DFT codes, it is possible to deterministically calculate the dispersive parameters of a channel by introducing some redundancies into the transmitting data.
Feature extraction is a fundamental operation of classification and pattern recognition. There ar... more Feature extraction is a fundamental operation of classification and pattern recognition. There are various strategies for one-and multi-dimensional feature extraction. The transform domain features are very effective when the patterns are characterized by their spectral properties. A wellknown successful example is the speech recognition. In this paper the feature extraction capability of discrete cosine transform (DCT), Walsh-Hadamard transform (WHT), discrete Hartley transform (DHT) and their sign transformations are investigated and compared for the recognition of two &mensional binary patterns. It is shown, in this paper, that the noise immunity of the transform based feature extraction is rather promising.
To overcome the difficulties of computation-intensive multimedia applications, the development gr... more To overcome the difficulties of computation-intensive multimedia applications, the development groups of major CPU manufactories, such as Intel""' and DigitalTM, have decided to include new instruction sets into their CPU families to increase their multimedia handling ability. The newly introduced instruction set is basically in a Single Instruction Multiple Data (SIMD) Stream operation type. For the practical purpose (e.g. the trade off between ihe complexity of hardware implementation and the soobtained performance improvement), they use a reduced SIMD instruction set instead of the full one. Taking Intel as an example, the new instruction set is composed of 57 operations called the MultiMedia extension (MMX) instruction set. Nowadays, how to fully utilize the power of the embedded instruction set for providing various multimedia applications becomes an interesting and important issue. In this paper, we demonstrate an efficient realization, based on the new MMX instruction set, of the block Inverse Discrete Cosine Transform (IDCT) and Motion Compensation (MC) which are kernel components of the block-based decoding standards, such as MPEG-1, H.261 and H.263. The convincing results show that: with the add of proper SIMD instruction set, the pure software solution for complicated multimedia applications (such as real-time MPEG video decoding) becomes feasible.
IEEE Transactions on Signal Processing, Mar 1, 1999
In this correspondence, an index permutation-based fast twodimensional discrete cosine transform ... more In this correspondence, an index permutation-based fast twodimensional discrete cosine transform (2-D DCT) algorithm is presented. It is shown that the N 2 N N 2 N N 2 N 2-D DCT, where N = 2 m N = 2 m N = 2 m , can be computed using only N N N 1-D DCT's and some post additions.
In this paper, a generalized output pruning algorithm for matrixvector multiplication is proposed... more In this paper, a generalized output pruning algorithm for matrixvector multiplication is proposed first. Then the application of the proposed pruning algorithm to compute pruning Discrete Cosine Transform (DCT) is addressed. It is shown that, for a given decomposition of the matrix of the transform kernel and the pruning pattern, the unnecessary operations for computing an output pruning DCT can be eliminated thoroughly by using the proposed algorithm. Consider the operation of a general matrix-vector multiplication of order N, say D,=A,,,xB N, and assume only partial multiplication outputs Dj (where 1 I j I N ) are required. It follows that we can speed up the aforecited computation by pruning the unnecessary operations. To reduce the computational complexity, we decompose the matrix A, , into
Huffman code [I] is the most well-known and widely used variable-length codes (VLCs) for compress... more Huffman code [I] is the most well-known and widely used variable-length codes (VLCs) for compressing various kinds of data. Conventional table lookup and/or bit-serial decoding techniques for realizing VLCs are inefficient in memory usage andor compntation'time. In this paper, we investigated some existing memory efficient VLC decoding methods [2, 3, 4, 51 and gave some comments and comparisons among them. We also proposed an SGH-Tree based data structure to implement Aggurwul's approach [5] such that a symbol can be decoded in constant time; moreover, the required memory is much smaller than that of the original approach.
In the above correspondence,' on page 1233, (6) is missing, on the right-hand side, a factor of T... more In the above correspondence,' on page 1233, (6) is missing, on the right-hand side, a factor of This missing phase factor eventually cancels and the final published pdf is correct as it stands. (Alternatively, one can absorb the missing phase into Q , whereupon Q is no longer Hermitian.
In this paper, an image authentication technique by embedding each image with a signature so as t... more In this paper, an image authentication technique by embedding each image with a signature so as to discourage unauthorized copying is proposed. The proposed technique could actually survive several kinds of image processing and the P E G lossy compression.
code for the node's label is transmitted. For the tree in Fig. la, its symbolic representation wo... more code for the node's label is transmitted. For the tree in Fig. la, its symbolic representation would be in the form shown in Fig. Ib. We rewrite it in a symbolic sequence as follows: 110110000000101000001000001011000000001000000 (1) Thus, 4N + 5 = 45 hits in total are required to transmit this quadtree structure by sending node labels in sequence. Next, we investigate the performance of the run-length coding algorithm applied to encode the quadtree structure and compare it with that of the above traditional label code. The first step of runlength coding is to compute the length of each run in which all elements have the same symbol and different ones from those in neighbouring runs. From eqn. 1, for example, this step results in a run-length sequence as shown in Fig. IC, which is as follows: 2127111515112816
In this work, three techniques for enhancing various chaos-based joint compression and encryption... more In this work, three techniques for enhancing various chaos-based joint compression and encryption (JCAE) schemes are proposed. They respectively improved the execution time, compression ratio, and estimation accuracy of three different chaos-based JCAE schemes. The first uses auxiliary data structures to significantly accelerate an existing chaos-based JCAE scheme. The second solves the problem of huge multidimensional lookup table overheads by sieving out a small number of important sub-tables. The third increases the accuracy of frequency distribution estimations, used for compressing streaming data, by weighting symbols in the plaintext stream according to their positions in the stream. Finally, two modified JCAE schemes leveraging the above three techniques are obtained, one applicable to static files and the other working for streaming data. Experimental results show that the proposed schemes do run faster and generate smaller files than existing JCAE schemes, which verified the effectiveness of the three newly proposed techniques.
Secure comparison is a fundamental problem in multiparty computation. There are two different par... more Secure comparison is a fundamental problem in multiparty computation. There are two different parties, each holding an l-bit integer, denoted by a and b, respectively. The goal of secure comparison is to compute the order relationship between a and b, say (a > b) ∈ {0, 1}, without revealing their inputs to any others. Since previous solutions based on homomorphic encryption need at least Ω(l) encryptions for each l-bit comparison, the total encryption time leads to a computational bottleneck for these protocols. This work presents a fast, semi-honest, secure comparison protocol based on the BFV encryption scheme. With its vector-like plaintext space, the number of required encryptions can be significantly reduced; actually, only six encryptions are needed for each comparison in our protocol. In other words, the proposed protocol can achieve the time complexity O(λ + l) for a given security parameter λ. As a result, 4096-bit integers can be securely compared within 12.08 ms, which is 280 times faster than the state-of-the-art homomorphic encryptionbased secure comparison protocol. Furthermore, we can compare k pairs of l • k -1 -bit integers with almost the same execution time as comparing l-bit integers and achieve higher throughput regardless of the compared integer size.
IEEE Transactions on Acoustics, Speech, and Signal Processing, Oct 1, 1988
(2.3) 2) The index i used in (2.7) and (2.12) should be replaced by 3) The term b(4, 16) used in ... more (2.3) 2) The index i used in (2.7) and (2.12) should be replaced by 3) The term b(4, 16) used in the lowest block of (2.13) should 4) Equation (A.16) should read n '. be replaced by b(9, 16). b ( m l , N ) = P ( m l ( l ) , N ) -P ( m 2 ( Nl ) , N ) . (A.
This article may be used for research, teaching and private study purposes. Any substantial or sy... more This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.
A class of neural network for constructing associative memories that learn the memory patterns as... more A class of neural network for constructing associative memories that learn the memory patterns as well as their neighbouring patterns is presented. The network is basically a layer of perceptrons with high-order polynomials as their discriminant functions. A learning algorithm IS proposed for the network to learn arbitrary bipolar patterns. The simulation results show that the associative memories implemented in this way achieve a set of desirable characteristics, namely high storage capacity, nearest convergence, and existence of a 'no decision' state which attracts indistinguishable inputs. Furthermore, it is also possible to shape the attraction basin of a memory pattern under any metria definition of distance.
With the growing popularity of cloud computing, it is convenient for data owners to outsource the... more With the growing popularity of cloud computing, it is convenient for data owners to outsource their data to a cloud server. By utilizing the massive storage and computational resources in cloud, data owners can also provide a platform for users to make query requests. However, due to the privacy concerns, sensitive data should be encrypted before outsourcing. In this work, a novel privacy preserving K-nearest neighbor (K-NN) search scheme over the encrypted outsourced cloud dataset is proposed. The problem is about letting the cloud server find K nearest points with respect to an encrypted query on the encrypted dataset, which was outsourced by data owners, and return the searched results to the querying user. Comparing with other existing methods, our approach leverages the resources of the cloud more by shifting most of the required computational loads, from data owners and query users, to the cloud server. In addition, there is no need for data owners to share their secret key with others. In a nutshell, in the proposed scheme, data points and user queries are encrypted attribute-wise and the entire search algorithm is performed in the encrypted domain; therefore, our approach not only preserves the data privacy and query privacy but also hides the data access pattern from the cloud server. Moreover, by using a tree structure, the proposed scheme could accomplish query requests in sub-liner time, according to our performance analysis. Finally, experimental results demonstrate the practicability and the efficiency of our method.
In theory, high key and high plaintext sensitivities are a must for a cryptosystem to resist the ... more In theory, high key and high plaintext sensitivities are a must for a cryptosystem to resist the chosen/known plaintext and the differential attacks. High plaintext sensitivity can be achieved by ensuring that each encrypted result is plaintext-dependent. In this work, we make detailed cryptanalysis on a published chaotic map-based image encryption system, where the encryption process is plaintext Image dependent. We show that some designing flaws make the published cryptosystem vulnerable to chosen-plaintext attack, and we then proposed an enhanced algorithm to overcome those flaws.
International Symposium/Conference on Music Information Retrieval, 2013
A musical medley is a piece of music that is composed of parts of existing pieces. Manually creat... more A musical medley is a piece of music that is composed of parts of existing pieces. Manually creating medley is time consuming because it is not easy to find out proper clips to put in succession and seamlessly connect them. In this work, we propose a framework for creating personalized music medleys from users' music collection. Unlike existing similar works in which only low-level features are used to select candidate clips and locate possible transition points among clips, we take song structures and song phrasing into account during medley creation. Inspired by the musical dice game, we treat the medley generation process as an audio version of musical dice game. That is, once the analysis on the songs of user collection has been done, the system is able to generate various medleys with different probabilities. This flexibility brings us the ability to create medleys according to the user-specified conditions, such as the medley structure or some must-use clips. The preliminary subjective evaluations showed that the proposed system is effective in selecting connectable clips that preserved chord progression structure. Besides, connecting the clips at phrase boundaries acquired more user preference than previous works did.
During medical treatment, personal privacy is involved and must be protected. Healthcare institut... more During medical treatment, personal privacy is involved and must be protected. Healthcare institutions have to keep medical images or health information secret unless they have permission from the data owner to disclose them. Reversible data hiding (RDH) is a technique that embeds metadata into an image and can be recovered without any distortion after the hidden data have been extracted. This work aims to develop a fully reversible two-bit embedding RDH algorithm with a large hiding capacity for medical images. Medical images can be partitioned into regions of interest (ROI) and regions of noninterest (RONI). ROI is informative with semantic meanings essential for clinical applications and diagnosis and cannot tolerate subtle changes. Therefore, we utilize histogram shifting and prediction error to embed metadata into RONI. In addition, our embedding algorithm minimizes the side effect to ROI as much as possible. To verify the effectiveness of the proposed approach, we benchmarked three types of medical images in DICOM format, namely Xray photography (X-ray), computer tomography (CT), and magnetic resonance imaging (MRI). Experimental results show that most of the hidden data have been embedded in RONI, and the performance achieves high capacity and leaves less visible distortion to ROIs.
Since a new discrete rectan lar wave transform (DRWT) has been propcsef there are various applica... more Since a new discrete rectan lar wave transform (DRWT) has been propcsef there are various applications which could be applied, such as real-time DFT implementation and ima e recognition. Although the DRWT possesses tfe property of easy-computation, there 19 no cost-effective algorithm for the implementation of DRWT. Most of the proposed architectures to implement this transform are systolic array, wavefront array, or algorithm based implementations with butterfly-like structures. Since these proposed architertiires have either the modular or high-throughput rate property, but not both; therefore, a same-geometry implementations B proposed in this paper. In this paper, a cost-effective algorithm for implementing DRWT is roposed, called the s a m T m e t r y DRWT. TKis newly proposed dgorit m provides a better procedure not merely in fast computation but also in the consideration of real implementation. In the sarne-geometry DRWT, power of two length is the cdnstraint. There are 21oaN-1 stages in the implementation of the DRWT with length N. Furthermore, logzN stages are identical and therefore only logzN st es demand circuit design in implementation phase.%he delay of each stage is merely one addition and the cells required to construct these stages are simple. Therefore, this algorithm is suitable for VLSI implementation.
The utilization of real-number DFT codes for channel equalization is studied in this letter. As s... more The utilization of real-number DFT codes for channel equalization is studied in this letter. As shown below, through real-number DFT codes, it is possible to deterministically calculate the dispersive parameters of a channel by introducing some redundancies into the transmitting data.
Feature extraction is a fundamental operation of classification and pattern recognition. There ar... more Feature extraction is a fundamental operation of classification and pattern recognition. There are various strategies for one-and multi-dimensional feature extraction. The transform domain features are very effective when the patterns are characterized by their spectral properties. A wellknown successful example is the speech recognition. In this paper the feature extraction capability of discrete cosine transform (DCT), Walsh-Hadamard transform (WHT), discrete Hartley transform (DHT) and their sign transformations are investigated and compared for the recognition of two &mensional binary patterns. It is shown, in this paper, that the noise immunity of the transform based feature extraction is rather promising.
To overcome the difficulties of computation-intensive multimedia applications, the development gr... more To overcome the difficulties of computation-intensive multimedia applications, the development groups of major CPU manufactories, such as Intel""' and DigitalTM, have decided to include new instruction sets into their CPU families to increase their multimedia handling ability. The newly introduced instruction set is basically in a Single Instruction Multiple Data (SIMD) Stream operation type. For the practical purpose (e.g. the trade off between ihe complexity of hardware implementation and the soobtained performance improvement), they use a reduced SIMD instruction set instead of the full one. Taking Intel as an example, the new instruction set is composed of 57 operations called the MultiMedia extension (MMX) instruction set. Nowadays, how to fully utilize the power of the embedded instruction set for providing various multimedia applications becomes an interesting and important issue. In this paper, we demonstrate an efficient realization, based on the new MMX instruction set, of the block Inverse Discrete Cosine Transform (IDCT) and Motion Compensation (MC) which are kernel components of the block-based decoding standards, such as MPEG-1, H.261 and H.263. The convincing results show that: with the add of proper SIMD instruction set, the pure software solution for complicated multimedia applications (such as real-time MPEG video decoding) becomes feasible.
IEEE Transactions on Signal Processing, Mar 1, 1999
In this correspondence, an index permutation-based fast twodimensional discrete cosine transform ... more In this correspondence, an index permutation-based fast twodimensional discrete cosine transform (2-D DCT) algorithm is presented. It is shown that the N 2 N N 2 N N 2 N 2-D DCT, where N = 2 m N = 2 m N = 2 m , can be computed using only N N N 1-D DCT's and some post additions.
In this paper, a generalized output pruning algorithm for matrixvector multiplication is proposed... more In this paper, a generalized output pruning algorithm for matrixvector multiplication is proposed first. Then the application of the proposed pruning algorithm to compute pruning Discrete Cosine Transform (DCT) is addressed. It is shown that, for a given decomposition of the matrix of the transform kernel and the pruning pattern, the unnecessary operations for computing an output pruning DCT can be eliminated thoroughly by using the proposed algorithm. Consider the operation of a general matrix-vector multiplication of order N, say D,=A,,,xB N, and assume only partial multiplication outputs Dj (where 1 I j I N ) are required. It follows that we can speed up the aforecited computation by pruning the unnecessary operations. To reduce the computational complexity, we decompose the matrix A, , into
Huffman code [I] is the most well-known and widely used variable-length codes (VLCs) for compress... more Huffman code [I] is the most well-known and widely used variable-length codes (VLCs) for compressing various kinds of data. Conventional table lookup and/or bit-serial decoding techniques for realizing VLCs are inefficient in memory usage andor compntation'time. In this paper, we investigated some existing memory efficient VLC decoding methods [2, 3, 4, 51 and gave some comments and comparisons among them. We also proposed an SGH-Tree based data structure to implement Aggurwul's approach [5] such that a symbol can be decoded in constant time; moreover, the required memory is much smaller than that of the original approach.
In the above correspondence,' on page 1233, (6) is missing, on the right-hand side, a factor of T... more In the above correspondence,' on page 1233, (6) is missing, on the right-hand side, a factor of This missing phase factor eventually cancels and the final published pdf is correct as it stands. (Alternatively, one can absorb the missing phase into Q , whereupon Q is no longer Hermitian.
In this paper, an image authentication technique by embedding each image with a signature so as t... more In this paper, an image authentication technique by embedding each image with a signature so as to discourage unauthorized copying is proposed. The proposed technique could actually survive several kinds of image processing and the P E G lossy compression.
code for the node's label is transmitted. For the tree in Fig. la, its symbolic representation wo... more code for the node's label is transmitted. For the tree in Fig. la, its symbolic representation would be in the form shown in Fig. Ib. We rewrite it in a symbolic sequence as follows: 110110000000101000001000001011000000001000000 (1) Thus, 4N + 5 = 45 hits in total are required to transmit this quadtree structure by sending node labels in sequence. Next, we investigate the performance of the run-length coding algorithm applied to encode the quadtree structure and compare it with that of the above traditional label code. The first step of runlength coding is to compute the length of each run in which all elements have the same symbol and different ones from those in neighbouring runs. From eqn. 1, for example, this step results in a run-length sequence as shown in Fig. IC, which is as follows: 2127111515112816
In this work, three techniques for enhancing various chaos-based joint compression and encryption... more In this work, three techniques for enhancing various chaos-based joint compression and encryption (JCAE) schemes are proposed. They respectively improved the execution time, compression ratio, and estimation accuracy of three different chaos-based JCAE schemes. The first uses auxiliary data structures to significantly accelerate an existing chaos-based JCAE scheme. The second solves the problem of huge multidimensional lookup table overheads by sieving out a small number of important sub-tables. The third increases the accuracy of frequency distribution estimations, used for compressing streaming data, by weighting symbols in the plaintext stream according to their positions in the stream. Finally, two modified JCAE schemes leveraging the above three techniques are obtained, one applicable to static files and the other working for streaming data. Experimental results show that the proposed schemes do run faster and generate smaller files than existing JCAE schemes, which verified the effectiveness of the three newly proposed techniques.
Secure comparison is a fundamental problem in multiparty computation. There are two different par... more Secure comparison is a fundamental problem in multiparty computation. There are two different parties, each holding an l-bit integer, denoted by a and b, respectively. The goal of secure comparison is to compute the order relationship between a and b, say (a > b) ∈ {0, 1}, without revealing their inputs to any others. Since previous solutions based on homomorphic encryption need at least Ω(l) encryptions for each l-bit comparison, the total encryption time leads to a computational bottleneck for these protocols. This work presents a fast, semi-honest, secure comparison protocol based on the BFV encryption scheme. With its vector-like plaintext space, the number of required encryptions can be significantly reduced; actually, only six encryptions are needed for each comparison in our protocol. In other words, the proposed protocol can achieve the time complexity O(λ + l) for a given security parameter λ. As a result, 4096-bit integers can be securely compared within 12.08 ms, which is 280 times faster than the state-of-the-art homomorphic encryptionbased secure comparison protocol. Furthermore, we can compare k pairs of l • k -1 -bit integers with almost the same execution time as comparing l-bit integers and achieve higher throughput regardless of the compared integer size.
IEEE Transactions on Acoustics, Speech, and Signal Processing, Oct 1, 1988
(2.3) 2) The index i used in (2.7) and (2.12) should be replaced by 3) The term b(4, 16) used in ... more (2.3) 2) The index i used in (2.7) and (2.12) should be replaced by 3) The term b(4, 16) used in the lowest block of (2.13) should 4) Equation (A.16) should read n '. be replaced by b(9, 16). b ( m l , N ) = P ( m l ( l ) , N ) -P ( m 2 ( Nl ) , N ) . (A.
This article may be used for research, teaching and private study purposes. Any substantial or sy... more This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.
A class of neural network for constructing associative memories that learn the memory patterns as... more A class of neural network for constructing associative memories that learn the memory patterns as well as their neighbouring patterns is presented. The network is basically a layer of perceptrons with high-order polynomials as their discriminant functions. A learning algorithm IS proposed for the network to learn arbitrary bipolar patterns. The simulation results show that the associative memories implemented in this way achieve a set of desirable characteristics, namely high storage capacity, nearest convergence, and existence of a 'no decision' state which attracts indistinguishable inputs. Furthermore, it is also possible to shape the attraction basin of a memory pattern under any metria definition of distance.
With the growing popularity of cloud computing, it is convenient for data owners to outsource the... more With the growing popularity of cloud computing, it is convenient for data owners to outsource their data to a cloud server. By utilizing the massive storage and computational resources in cloud, data owners can also provide a platform for users to make query requests. However, due to the privacy concerns, sensitive data should be encrypted before outsourcing. In this work, a novel privacy preserving K-nearest neighbor (K-NN) search scheme over the encrypted outsourced cloud dataset is proposed. The problem is about letting the cloud server find K nearest points with respect to an encrypted query on the encrypted dataset, which was outsourced by data owners, and return the searched results to the querying user. Comparing with other existing methods, our approach leverages the resources of the cloud more by shifting most of the required computational loads, from data owners and query users, to the cloud server. In addition, there is no need for data owners to share their secret key with others. In a nutshell, in the proposed scheme, data points and user queries are encrypted attribute-wise and the entire search algorithm is performed in the encrypted domain; therefore, our approach not only preserves the data privacy and query privacy but also hides the data access pattern from the cloud server. Moreover, by using a tree structure, the proposed scheme could accomplish query requests in sub-liner time, according to our performance analysis. Finally, experimental results demonstrate the practicability and the efficiency of our method.
In theory, high key and high plaintext sensitivities are a must for a cryptosystem to resist the ... more In theory, high key and high plaintext sensitivities are a must for a cryptosystem to resist the chosen/known plaintext and the differential attacks. High plaintext sensitivity can be achieved by ensuring that each encrypted result is plaintext-dependent. In this work, we make detailed cryptanalysis on a published chaotic map-based image encryption system, where the encryption process is plaintext Image dependent. We show that some designing flaws make the published cryptosystem vulnerable to chosen-plaintext attack, and we then proposed an enhanced algorithm to overcome those flaws.
Uploads
Papers by Ja-Ling Wu