This paper presents an error-resilient MPEG-4 video communication system. The system comprises an... more This paper presents an error-resilient MPEG-4 video communication system. The system comprises an error-resilient encoder, an adaptive error-resilient transcoder, an error-resilient decoder and the PLBP encoderidecoder. It has been evaluated under a configurable network error simulator. In addition, the system integrates the error resilience tools defined in MPEG-4 visual standard, namely Video Packet Structure, Header Extension Code (HEC) and NEWPRED, as well as three proposed novel error-resilient techniques, namely Prioritized-Level Bitstream Protection (PLBP), Protected Data Partitioning (PDP) and Restorative Intra Adoption (RIA). In the decoder, three error concealment techniques are implemented. They are temporal replacement, motion replacement, and predictive motion replacement. The decoder also implements an error detection mechanism that has high sensitivity to catch errors. Finally, the performance of the hybrid error-resilient system has been evaluated. The experimental results show that under the error-prone networks, acceptable video quality has been successfully retained without incurring much overhead.
In this paper, we propose an encryption scheme for MPEG video, which uses a permuted HufFman tabl... more In this paper, we propose an encryption scheme for MPEG video, which uses a permuted HufFman table to encode the input symbols, and then rotates and XORs (exclusive or) the encoded bit stream. This scheme adds very little computational overhead to the MPEG video compression process, and therefore the implementation is fast enough to meet the real-time requirement of MPEG video applications. In ow analysis, it is also robust to both plaintext and ciphertext attacks. In addition, OUT scheme can be applied to all MPEG compression standards.
EURASIP Journal on Advances in Signal Processing, Oct 8, 2008
Massively multiplayer online role playing games (MMORPGs) have become extremely popular among net... more Massively multiplayer online role playing games (MMORPGs) have become extremely popular among network gamers. Despite their success, one of MMORPG's greatest challenges is the increasing use of game bots, that is, autoplaying game clients. The use of game bots is considered unsportsmanlike and is therefore forbidden. To keep games in order, game police, played by actual human players, often patrol game zones and question suspicious players. This practice, however, is labor-intensive and ineffective. To address this problem, we analyze the traffic generated by human players versus game bots and propose general solutions to identify game bots. Taking Ragnarok Online as our subject, we study the traffic generated by human players and game bots. We find that their traffic is distinguishable by 1) the regularity in the release time of client commands, 2) the trend and magnitude of traffic burstiness in multiple time scales, and 3) the sensitivity to different network conditions. Based on these findings, we propose four strategies and two ensemble schemes to identify bots. Finally, we discuss the robustness of the proposed methods against countermeasures of bot developers, and consider a number of possible ways to manage the increasingly serious bot problem.
Massively multiplayer online role playing games (MMORPGs) have become extremely popular among net... more Massively multiplayer online role playing games (MMORPGs) have become extremely popular among network gamers. Despite their success, one of MMORPG's greatest challenges is the increasing use of game bots, that is, autoplaying game clients. The use of game bots is considered unsportsmanlike and is therefore forbidden. To keep games in order, game police, played by actual human players, often patrol game zones and question suspicious players. This practice, however, is labor-intensive and ineffective. To address this problem, we analyze the traffic generated by human players versus game bots and propose general solutions to identify game bots. Taking Ragnarok Online as our subject, we study the traffic generated by human players and game bots. We find that their traffic is distinguishable by 1) the regularity in the release time of client commands, 2) the trend and magnitude of traffic burstiness in multiple time scales, and 3) the sensitivity to different network conditions. Based on these findings, we propose four strategies and two ensemble schemes to identify bots. Finally, we discuss the robustness of the proposed methods against countermeasures of bot developers, and consider a number of possible ways to manage the increasingly serious bot problem.
In this paper, we propose a novel system that can real-time capture and compress the full screen ... more In this paper, we propose a novel system that can real-time capture and compress the full screen of PC into a video clip. It is real-time in that it can capture to 30 frames per second under the resolution of 1600X 1200 with True Color. One application of this system is to produce a digital presentation clip for instruction or tutorial. Moreover, as the video clips can be streamed over Internet or Intranet, they can be used for remote education or training. We believe this approach is clearer and more efficient than conventional text manual or handbook. As our system only captures the differences of successive snapshots instead of every single screen, it is more efficient and produced more compact clips than other existing systems. In addition, the compression algorithm adopted in our system is also described in this paper.
With the popularity of social media nowadays, tons of photos are uploaded everyday. To understand... more With the popularity of social media nowadays, tons of photos are uploaded everyday. To understand the image content, image classification becomes a very essential technique for plenty of applications (e.g., object detection, image caption generation). Convolutional Neural Network (CNN) has been shown as the state-of-the-art approach for image classification. However, one of the characteristics in social media photos is that they are often applied with photo filters, especially on Instagram. We find that prior works do not aware of this trend in social media photos and fail on filtered images. Thus, we propose a novel CNN architecture that utilizes the power of pairwise constraint by combining Siamese network and the proposed adaptive margin contrastive loss with our discriminative pair sampling method to solve the problem of filter bias. To the best of our knowledge, this is the first work to tackle filter bias on CNN and achieve stateof-the-art performance on a filtered subset of ILSVRC2012.
Removing shadows in document images enhances both the visual quality and readability of digital c... more Removing shadows in document images enhances both the visual quality and readability of digital copies of documents. Most existing shadow removal algorithms for document images use hand-crafted heuristics and are often not robust to documents with different characteristics. This paper proposes the Background Estimation Document Shadow Removal Network (BEDSR-Net), the first deep network specifically designed for document image shadow removal. For taking advantage of specific properties of document images, a background estimation module is designed for extracting the global background color of the document. During the process of estimating the background color, the module also learns information about the spatial distribution of background and non-background pixels. We encode such information into an attention map. With the estimated global background color and attention map, the shadow removal network can better recover the shadow-free image. We also show that the model trained on synthetic images remains effective for real photos, and provide a large set of synthetic shadow images of documents along with their corresponding shadow-free images and shadow masks. Extensive quantitative and qualitative experiments on several benchmarks show that the BEDSR-Net outperforms existing methods in enhancing both the visual quality and readability of document images.
We show how to transform the B-spline curve and surface fitting problems into suffix computations... more We show how to transform the B-spline curve and surface fitting problems into suffix computations of continued fractions. Then a parallel substitution scheme is introduced to compute the suffix values on a newly proposed mesh-of-unshuffle network. The derived parallel algorithm allows the curve interpolation through n points to be solved in O(logn) time using @(n/ logn) processors and allows the surface interpolation through m x n points to be solved in O(log mlog n) time using 0 (mn/(log m log n)) processors. Both interpolation algorithms are cost-optimal for their respective problems. Besides, the surface fitting problem can be even faster solved in O(log m + log n) time if o(mn) processors are used in the network.
finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neit... more finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neither algorithm fails in compacting the examples. However, since this is still an increase over no foresight, it is better to use the incremental foresight algorithm only when the DPS fails.
Direct chaining is a popular and efficient class of hashing algorithms. In this paper we study op... more Direct chaining is a popular and efficient class of hashing algorithms. In this paper we study optimum algorithms among direct chaining methods, under the restrictions that the records in the hash table are not moved after they are inserted, that for each chain the relative ordering of the records in the chain does not change after more insertions, and that only one link field is used per table slot. The varied-insertion coalesced hashing method (VICH), which is proposed and analyzed in [CV84], is conjectured to be optimum among all direct chaining algorithms in this class. We give strong evidence in favor of the conjecture hy showing that VICH is optimum under fairly general conditions. Key words, analysis of algorithms, searching, information retrieval, hashing, coalesced hashing, data structures, optimality 1. Introduction. There are many classes of hashing algorithms in use today" separate chaining, coalesced hashing, linear probing, double hashing, and quadratic probing, to name a few. (More details can be found in [Knu73].) Comparisons between hashing algorithms in different classes are often difficult, because each class has its own assumptions, storage requirements, and tradeoffs. For example, some hashing algorithms (as in [Bre73]) do extra work during insertion in order to speed up later searches. In some applications, the preferred class of hashing methods is determined
International Journal of Foundations of Computer Science, 1994
We consider heap-ordered trees in this paper. The main theorems in this paper show that the avera... more We consider heap-ordered trees in this paper. The main theorems in this paper show that the average and the variance of the altitude (or level) of the nodes in a random n-node heap-ordered tree are [Formula: see text] and [Formula: see text], respectively where γ=0.5772156649… is the Euler’s constant.
We show that the average internal path length of the binary representation of a random n-node hea... more We show that the average internal path length of the binary representation of a random n-node heap-ordered tree is asymptotically n&G-2n + O(fi).
finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neit... more finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neither algorithm fails in compacting the examples. However, since this is still an increase over no foresight, it is better to use the incremental foresight algorithm only when the DPS fails.
Multimedia applications have become extremely popular among daily life. Despite the availability ... more Multimedia applications have become extremely popular among daily life. Despite the availability of the untold multimedia application, the desire for novel multimedia applications, such as digital learning applications, digital TV and multimedia home entertainment, is desperately ...
This paper presents an error-resilient MPEG-4 video communication system. The system comprises an... more This paper presents an error-resilient MPEG-4 video communication system. The system comprises an error-resilient encoder, an adaptive error-resilient transcoder, an error-resilient decoder and the PLBP encoderidecoder. It has been evaluated under a configurable network error simulator. In addition, the system integrates the error resilience tools defined in MPEG-4 visual standard, namely Video Packet Structure, Header Extension Code (HEC) and NEWPRED, as well as three proposed novel error-resilient techniques, namely Prioritized-Level Bitstream Protection (PLBP), Protected Data Partitioning (PDP) and Restorative Intra Adoption (RIA). In the decoder, three error concealment techniques are implemented. They are temporal replacement, motion replacement, and predictive motion replacement. The decoder also implements an error detection mechanism that has high sensitivity to catch errors. Finally, the performance of the hybrid error-resilient system has been evaluated. The experimental results show that under the error-prone networks, acceptable video quality has been successfully retained without incurring much overhead.
In this paper, we propose an encryption scheme for MPEG video, which uses a permuted HufFman tabl... more In this paper, we propose an encryption scheme for MPEG video, which uses a permuted HufFman table to encode the input symbols, and then rotates and XORs (exclusive or) the encoded bit stream. This scheme adds very little computational overhead to the MPEG video compression process, and therefore the implementation is fast enough to meet the real-time requirement of MPEG video applications. In ow analysis, it is also robust to both plaintext and ciphertext attacks. In addition, OUT scheme can be applied to all MPEG compression standards.
EURASIP Journal on Advances in Signal Processing, Oct 8, 2008
Massively multiplayer online role playing games (MMORPGs) have become extremely popular among net... more Massively multiplayer online role playing games (MMORPGs) have become extremely popular among network gamers. Despite their success, one of MMORPG's greatest challenges is the increasing use of game bots, that is, autoplaying game clients. The use of game bots is considered unsportsmanlike and is therefore forbidden. To keep games in order, game police, played by actual human players, often patrol game zones and question suspicious players. This practice, however, is labor-intensive and ineffective. To address this problem, we analyze the traffic generated by human players versus game bots and propose general solutions to identify game bots. Taking Ragnarok Online as our subject, we study the traffic generated by human players and game bots. We find that their traffic is distinguishable by 1) the regularity in the release time of client commands, 2) the trend and magnitude of traffic burstiness in multiple time scales, and 3) the sensitivity to different network conditions. Based on these findings, we propose four strategies and two ensemble schemes to identify bots. Finally, we discuss the robustness of the proposed methods against countermeasures of bot developers, and consider a number of possible ways to manage the increasingly serious bot problem.
Massively multiplayer online role playing games (MMORPGs) have become extremely popular among net... more Massively multiplayer online role playing games (MMORPGs) have become extremely popular among network gamers. Despite their success, one of MMORPG's greatest challenges is the increasing use of game bots, that is, autoplaying game clients. The use of game bots is considered unsportsmanlike and is therefore forbidden. To keep games in order, game police, played by actual human players, often patrol game zones and question suspicious players. This practice, however, is labor-intensive and ineffective. To address this problem, we analyze the traffic generated by human players versus game bots and propose general solutions to identify game bots. Taking Ragnarok Online as our subject, we study the traffic generated by human players and game bots. We find that their traffic is distinguishable by 1) the regularity in the release time of client commands, 2) the trend and magnitude of traffic burstiness in multiple time scales, and 3) the sensitivity to different network conditions. Based on these findings, we propose four strategies and two ensemble schemes to identify bots. Finally, we discuss the robustness of the proposed methods against countermeasures of bot developers, and consider a number of possible ways to manage the increasingly serious bot problem.
In this paper, we propose a novel system that can real-time capture and compress the full screen ... more In this paper, we propose a novel system that can real-time capture and compress the full screen of PC into a video clip. It is real-time in that it can capture to 30 frames per second under the resolution of 1600X 1200 with True Color. One application of this system is to produce a digital presentation clip for instruction or tutorial. Moreover, as the video clips can be streamed over Internet or Intranet, they can be used for remote education or training. We believe this approach is clearer and more efficient than conventional text manual or handbook. As our system only captures the differences of successive snapshots instead of every single screen, it is more efficient and produced more compact clips than other existing systems. In addition, the compression algorithm adopted in our system is also described in this paper.
With the popularity of social media nowadays, tons of photos are uploaded everyday. To understand... more With the popularity of social media nowadays, tons of photos are uploaded everyday. To understand the image content, image classification becomes a very essential technique for plenty of applications (e.g., object detection, image caption generation). Convolutional Neural Network (CNN) has been shown as the state-of-the-art approach for image classification. However, one of the characteristics in social media photos is that they are often applied with photo filters, especially on Instagram. We find that prior works do not aware of this trend in social media photos and fail on filtered images. Thus, we propose a novel CNN architecture that utilizes the power of pairwise constraint by combining Siamese network and the proposed adaptive margin contrastive loss with our discriminative pair sampling method to solve the problem of filter bias. To the best of our knowledge, this is the first work to tackle filter bias on CNN and achieve stateof-the-art performance on a filtered subset of ILSVRC2012.
Removing shadows in document images enhances both the visual quality and readability of digital c... more Removing shadows in document images enhances both the visual quality and readability of digital copies of documents. Most existing shadow removal algorithms for document images use hand-crafted heuristics and are often not robust to documents with different characteristics. This paper proposes the Background Estimation Document Shadow Removal Network (BEDSR-Net), the first deep network specifically designed for document image shadow removal. For taking advantage of specific properties of document images, a background estimation module is designed for extracting the global background color of the document. During the process of estimating the background color, the module also learns information about the spatial distribution of background and non-background pixels. We encode such information into an attention map. With the estimated global background color and attention map, the shadow removal network can better recover the shadow-free image. We also show that the model trained on synthetic images remains effective for real photos, and provide a large set of synthetic shadow images of documents along with their corresponding shadow-free images and shadow masks. Extensive quantitative and qualitative experiments on several benchmarks show that the BEDSR-Net outperforms existing methods in enhancing both the visual quality and readability of document images.
We show how to transform the B-spline curve and surface fitting problems into suffix computations... more We show how to transform the B-spline curve and surface fitting problems into suffix computations of continued fractions. Then a parallel substitution scheme is introduced to compute the suffix values on a newly proposed mesh-of-unshuffle network. The derived parallel algorithm allows the curve interpolation through n points to be solved in O(logn) time using @(n/ logn) processors and allows the surface interpolation through m x n points to be solved in O(log mlog n) time using 0 (mn/(log m log n)) processors. Both interpolation algorithms are cost-optimal for their respective problems. Besides, the surface fitting problem can be even faster solved in O(log m + log n) time if o(mn) processors are used in the network.
finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neit... more finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neither algorithm fails in compacting the examples. However, since this is still an increase over no foresight, it is better to use the incremental foresight algorithm only when the DPS fails.
Direct chaining is a popular and efficient class of hashing algorithms. In this paper we study op... more Direct chaining is a popular and efficient class of hashing algorithms. In this paper we study optimum algorithms among direct chaining methods, under the restrictions that the records in the hash table are not moved after they are inserted, that for each chain the relative ordering of the records in the chain does not change after more insertions, and that only one link field is used per table slot. The varied-insertion coalesced hashing method (VICH), which is proposed and analyzed in [CV84], is conjectured to be optimum among all direct chaining algorithms in this class. We give strong evidence in favor of the conjecture hy showing that VICH is optimum under fairly general conditions. Key words, analysis of algorithms, searching, information retrieval, hashing, coalesced hashing, data structures, optimality 1. Introduction. There are many classes of hashing algorithms in use today" separate chaining, coalesced hashing, linear probing, double hashing, and quadratic probing, to name a few. (More details can be found in [Knu73].) Comparisons between hashing algorithms in different classes are often difficult, because each class has its own assumptions, storage requirements, and tradeoffs. For example, some hashing algorithms (as in [Bre73]) do extra work during insertion in order to speed up later searches. In some applications, the preferred class of hashing methods is determined
International Journal of Foundations of Computer Science, 1994
We consider heap-ordered trees in this paper. The main theorems in this paper show that the avera... more We consider heap-ordered trees in this paper. The main theorems in this paper show that the average and the variance of the altitude (or level) of the nodes in a random n-node heap-ordered tree are [Formula: see text] and [Formula: see text], respectively where γ=0.5772156649… is the Euler’s constant.
We show that the average internal path length of the binary representation of a random n-node hea... more We show that the average internal path length of the binary representation of a random n-node heap-ordered tree is asymptotically n&G-2n + O(fi).
finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neit... more finite maximum arcs. Overall, the incremental algorithm saves around 48% of excess time, and neither algorithm fails in compacting the examples. However, since this is still an increase over no foresight, it is better to use the incremental foresight algorithm only when the DPS fails.
Multimedia applications have become extremely popular among daily life. Despite the availability ... more Multimedia applications have become extremely popular among daily life. Despite the availability of the untold multimedia application, the desire for novel multimedia applications, such as digital learning applications, digital TV and multimedia home entertainment, is desperately ...
Uploads
Papers by Wen-chin Chen