Search

In the field of unsupervised person re-identification (Re-ID), mainstream methods adopt cluster algorithm to generate pseudo labels for training. Despite the effectiveness, the cluster algorithm generates noisy labels, which are retained in further model updating and hinder higher performance. To solve this problem, we propose a Two-branch Asymmetric Model with Alternately Clustering. Specifically, the designed Alternately Clustering (AC) strategy leverages a two-branch model to cluster different pseudo labels for each branch, which prevents the continuous existence of identical noisy labels. To establish a mapping between the two label sets, pseudo label mapping constraint (PLMC) module is devised, which helps retain reliable pseudo labels. Our method improves the quality of generated pseudo labels by keeping noisy labels changing and retaining the reliable ones. Experimental results demonstrate that our proposed method outperforms the state-of-the-art methods.

Detecting temporal actions in long and untrimmed videos is a challenging and important field in computer vision. Generating high-quality proposals is a key step in temporal action detection. A high-quality proposal usually contains two main characteristics. One is the temporal overlaps between proposals and action instances should be as large as possible. The another one is the number of generated proposals should be as few as possible. Inspired by the similarity comparison in face recognition and the similarity of action in same action segment, we design a module to compare the similarity for visual features extracted from visual feature encoder. We find out time points where the similarity of features changes shapely to generate candidate proposals. Then, we train a classifier to evaluate the candidate proposals whether contains or not contains action instances. The experiments suggest that our method outperforms other temporal action proposal generation methods in THUMOS-14 dataset and ActivityNet-v1.3 dataset. In addition, our method still outperforms other methods when using different visual features extracted from different networks.

The past few years in the fields of Person Re-Identification (RE-ID) have seen attention mechanism receives enormous interest as it has superior performance in obtaining discriminative feature representations. However, a wide range of state-of-the-art RE-ID attention models only focus on one-dimensional attention design method, e.g. spatial attention and channels attention, hence the produced attention maps are neither detailed enough nor discriminative enough to capture complicated interactions of visual parts. Developing multi-scale attention mechanism for RE-ID, an under-studied approach, becomes a practicable method to overcome this deficiency. Toward this goal, we propose a Multiscale Omnibearing Attention Networks (MOAN) for RE-ID which is capable of utilizing the complex fusion information acquired from the multiscale attention mechanism with features being more representative. Specifically, MOAN takes full advantage of multi-sized convolution filters to obtain discriminative holistic and local feature maps, and adaptively conducts feature information augmentation by introducing an Omnibearing Attention (OA) module. Through the OA module, spatial attention and channel attention are integrated together in a unique way where they work in a complementary way. To sum up, MOAN not only inherits the merit of two kinds of attention mechanism but also performs well in extracting comprehensive feature information. Furthermore, taking into account the robustness of model performance, we formulate a Random Drop (RD) Function to facilitate training MOAN and further increase the diversity of training model for adaptation. Furthermore, to achieve end-to-end training, we utilize trainable parameters to take place of initial fixed parameters, and the model performance is experimentally promoted. Extensive experiments have been carried out on the four mainstream RE-ID datasets. As the result shows, our method with re-ranking achieves rank-1 accuracy of 92.29% on CUHK03-NP, 97.45% on Market-1501, 93.81% on DukeMTMC-reID and 81.53% on MSMT17-V2, outperforming the state-of-the-art methods and confirming the effectiveness of our method.

Image‐to‐image translation is a class of vision and graphics problems where the goal is to learn the mapping between input images and output images. However, due to the unstable training and limited training samples, many existing GAN‐based works have difficulty in producing photo‐realistic images. Herein, dual‐directional generative adversarial networks are proposed, which consist of four adversarial networks, to produce images of high perceptual quality. In this framework, self‐reconstruction strategy is used to construct auxiliary sub‐networks, which impose more effective constraints on encoder‐generator pairs. Using this idea, this model can increase the use ratio of paired data conditioned on the same dataset and obtain well‐trained encoder‐generator pairs with the help of the proposed cross‐network skip connections. Moreover, the proposed framework not only produces realistic images but also addresses the problem where condition GAN produces sharp images containing many small, hallucinated objects. Training on multiple supervised datasets, convincing evidences are shown to prove that this model can achieve compelling results by latently learning a common feature representation. Qualitative and quantitative comparisons against other methods, demonstrate the effectiveness and superiority of the method.

RGB‐Infrared person re‐identification (RGB‐IR Re‐ID) is a task aiming to retrieve and match person images between RGB images and IR images. Since most surveillance cameras capture RGB images during the day and IR images at night, RGB‐IR Re‐ID is helpful when checking day and night surveillance for criminal investigations. Previous related work often only extracts sharable and identity‐related features in images for identification. Few researches specifically extract and make use of features that do not have the ability to distinguish identity, e.g. identity‐unrelated features derived from background and modality. In this Letter, we propose a novel and concise RGB‐IR Re‐ID network named two‐way constraint network (TWCN). Compared with traditional Re‐ID networks, TWCN not only extracts and utilises identity‐related features but also novelly makes full use of identity‐unrelated features to improve the accuracy of the experiment. TWCN uses a reverse‐triplet loss to extract identity‐unrelated features, and proposes an orthogonal constraint to remove identity‐unrelated information from identity‐related features, which improves the purity of identity‐related features. In addition, a correlation coefficient synergy and central clustering (CCSCC) loss is introduced into TWCN to extract identity‐related features effectively. Extensive experiments have been conducted to prove our method is effective.

The research of part-based methods has been proven as an effective way in person re-identification (Re-ID) task. However, in existing part-based Re-ID methods, the informative interactions and potential associations among parts are neglected, which demands further study. To fill this gap, we propose a novel Part-relation-aware Feature Fusion Network (PFFN) which achieves a part-level feature fusion and enhances the discrimination of part features by fully employing helpful information from associations among parts. More specifically, a Dual-stage Attention (DA) module, consisting of spatial and part-based channel attention, is proposed to exploit complementary benefits of two kinds of attention information, thereby facilitating model with learning more discriminative features. Furthermore, Part-relation Exploitation (PE) module is proposed to learn relation-aware part features where correlative information among parts are fully employed, thereby bringing a noticeable improvement in performance. Extensive experiments are conducted on four mainstream Re-ID datasets to verify the superiority of PFFN. Compared with baseline model, PFFN has gained rank-1 accuracy improvement of 18.2% on MSMT17-v2, 12.1% on the CUHK03-Labeled, 5.6% on DukeMTMC-reid and 1.3% on Market1501, compellingly validating its effectiveness.

In this work, we propose a Three-Dimensional Transmissible Attention Network (3DTANet) for Person Re-Identification, which can transmit the attention information from layer to layer and attend to the person image from a three-dimensional perspective. Main contributions of the 3DTANet are: (i) A novel Transmissible Attention (TA) mechanism is introduced, which can transfer attention information between convolution layers. Different from traditional attention mechanism, not only can it convey accumulated attention information layer by layer but also guide the network to retain holistic attention information. (ii) We propose a Three-Dimension Attention (3DA) mechanism, which is capable of extracting a three-dimensional attention map. While previous researches on image attention mechanism extracts channel or spatial attention information separately, 3DA mechanism pays attention to channel and spatial information simultaneously, thereby making them play better complementary role in attention extraction. (iii) A new loss function named L2-norm Multi-labels Loss (L2ML) is applied to acquire higher recognition accuracy calculated by multi labels of same ID and corresponding feature representation. Quite different from the common loss functions, L2-norm Multi-labels Loss is specifically good at optimizing feature distance. In brief, 3DTANet gains two-fold benefit toward higher accuracy. For one thing, the attention information is informative and can be transmitted, feature being more representative. For another, our model is computationally lightweight and can be easily applied to real scenarios. We extensively conduct experiments on four Person Re-Identification benchmark datasets. Our model achieves rank-1 accuracy of 87.50% on CUHK03, 96.23% on Market-1501, 92.50% on DukeMTMC-reID and 76.60% on MSMT17-V2 respectively. The results confirm that the 3DTANet can extract more representative features and attain a higher recognition accuracy, outperforming the state-of-the-art methods.

Person Re-identification (ReID) aims to match people across non-overlapping camera views in a public space, which is usually regarded as an image retrieval problem to match query images with pedestrian images in the gallery. It is challenging since many difficulties exist such as pose misalignments, occlusions, similar appearance when detecting people. Existing researches on ReID mainly focus on two major problems: representation learning and metric learning. In this paper, we target at learning discriminative representations and make two contributions in total. ( $i$ ) We propose a novel architecture named Deeply Associative Two-stage Representations Learning (DATRL). It contains the global re-initialization stage and fully-perceptual classification stage employing two identical CNNs associatively at the same time. On the global stage, we take on the backbone of one deep CNN e.g., dozens of layers in the front of Resnet-50 as a normal re-initialization subnetwork. Meanwhile, we apply our own proposed 3D-transpose technique into the backbone of the other CNN to form the 3D-transpose re-initialization subnetwork. The fully-perceptual stage is actually made up of the leftover layers of the original CNNs. On this stage, we take both the global representations learned at multiple hierarchies and the local representations uniformly-partitioned on the highest conv-layer into consideration, and then optimizing them separately for classification. ( $ii$ ) We introduce a new joint loss function in which our proposed Labels Interval Extension loss (LIEL) and Group loss (GL) are combined to enhance the performance of gradient decent as well as increasing the distances between image features with different identities. We apply the above DATRL, LIEL and GL to ReID thus obtaining DATRL-ReID. Experimental results on four datasets CUHK03, Market-1501, DukeMTMC-reID and MSMT17-V2 demonstrate that DATRL-ReID shows excellent performance in improving recognition accuracy and is superior to state-of-the-art methods.

An output-capacitorless low-dropout regulator (OCL-LDO) using a dual-active feedback frequency compensation (DAFFC) scheme with both transient and stability enhancement has been presented in this paper. The DAFFC scheme consists of two parallel active feedback paths, which creates two pole-zero pairs to effectively enhance the stability and transient response for the proposed OCL-LDO. Compared to the conventional single-path active-feedback frequency compensation method, the proposed DAFFC technique has provided one more design freedom with one more active feedback loop deployed and has been proved to be capable of obtaining better compensation effects with the same capacitor budget. Besides, the induced extra ac currents by the two active feedback loops have also enhanced the transient response of the proposed OCL-LDO. To substantiate the proposed DAFFC, a telescopic cascode output stage for error amplifier, and two on-chip compensation capacitors (5 and 1 pF, respectively) are needed. The proposed OCL-LDO has been implemented in 65-nm CMOS technology and the active chip area is 0.0105 mm2. The output voltage is 0.8 V, and the minimum input voltage is 0.95 V at 100-mA loading current. The proposed OCL-LDO can work stably in a load range of 0 to 100 mA with 14-μA quiescent current.

RISC-V is an open source instruction set architecture (ISA) based on the principles of Reduced Instruction Set Computing (RISC). High-level synthesis (HLS) can automatically synthesize high-level specifications (such as in C or C++) into low-level RTL specifications for efficient implementation in application-specific integrated circuits (ASIC) or field programmable gate arrays (FPGA). This article proposes a method to implement CPU based on HLS. Based on the RISC-V architecture, the processor is designed using C language and synthesized into RTL code through HLS. This method greatly improve the speed of design and reduce the manpower required for design. In addition, by using Universal Verification Methodology (UVM) to build a verification platform to verify the RTL code, the synthesized RTL code is reliably and fully verified.

Attention mechanisms are widely used in re-identification (reID) tasks, but few attention-based architectures have considered integrating local features with their global dependencies, that is the previous works do not model the semantic interdependencies in both spatial dimension and channel dimension. Intuitively, for person feature representations, it is important to model the interdependencies between human body semantics. In this Letter, the authors proposed a dual semantic interdependencies attention module to capture semantic interdependencies in both spatial dimension and channel dimension simultaneously. Besides, they designed a deep supervision branch to directly guide the training of the attention modules and innovatively introduce a channels random dropping mechanism in the training phase to promote the attention modules to capture more discriminative information. Extensive experimental results show that the network merging the above strategies achieves state-of-the-art results on the mainstream reID data sets.

Recently, transformer has been applied to the image caption model, in which the convolutional neural network and the transformer encoder act as the image encoder of the model, and the transformer decoder acts as the decoder of the model. However, transformer may suffer from the interference of non-critical objects of a scene and meet with difficulty to fully capture image information due to its self-attention mechanism's dense characteristics. In this Letter, in order to address this issue, the authors propose a novel transformer model with decreasing attention gates and attention fusion module. Specifically, they firstly use attention gate to force transformer to overcome the interference of non-critical objects and capture objects information more efficiently via truncating all the attention weights that smaller than gate threshold. Secondly, through inheriting attentional matrix from the previous layer of each network layer, the attention fusion module enables each network layer to consider other objects without losing the most critical ones. Their method is evaluated using the benchmark Microsoft COCO dataset and achieves better performance compared to the state-of-the-art methods.

Collaborative representation based classification (CRC) has been widely used and shown good performance in face recognition (FR). Afterwards, hierarchical representation based classification has recently been proposed and aims to enhance the classification performance of the CRC method. However, these methods highly depend on the over-complete dictionary comprised of sufficient training samples, and cannot be directly applied for single-sample FR. In this study, the authors propose a novel CRC-based FR framework to address this issue, which is named hierarchical extended collaborative representation based classification (HECRC). Firstly, they integrate hierarchical representation based model with low-rank constrained variation dictionaries. Secondly, they select training samples that are the nearest neighbours of test images to obtain a more discriminative training dictionary, where an adaptive scheme is introduced to select proper samples automatically instead of setting a predefined number in traditional methods. Finally, the refined training dictionary and the learned variation dictionaries are jointly utilised to represent the test sample. Moreover, they combined the proposed HECRC with deep features to further improve the recognition rate. Experiments have been conducted on the AR and FERET datasets, and the results show that the proposed method has a substantial improvement over existing algorithms for single-sample FR.

Temporal action detection, as a branch of video analysis, aims to locate the time points when the actions start and end, and classify the actions occurred in videos into correct categories. Generating high-quality proposals is a key step in temporal action detection task. In this paper, we introduce a novel network, named multi-scale proposal regression network (MPRN), for temporal action proposal generation. First, we take encoding visual features as input and predict action scores for time points, in order to group them to generate rough proposals. Then, we regress the proposal's boundaries to obtain more precise proposals via our multi-scale proposal regression network. Compared with SSN and TURN, our multi-scale regression segments are characterized by flexible boundaries. Experiments show that 1) Our method is better than other proposal generation methods on THUMOS-14 dataset and ActivityNet-v1.3 dataset. 2) The effectiveness of our method is due to its own architecture, not the selection of visual feature encoders. 3) Our proposal generation method can generate temporal proposals for unseen action classes, which shows the good generalization ability of our proposal generation method.

Early compressor trees based on carry-save adders and single-column parallel counters show good performance in ASIC design, but do not adapt well to modern field-programmable gate arrays (FPGAs). Recently, compressor trees built from generalized parallel counters (GPCs) were synthesized on FPGAs to address this issue. Despite the improved timing performance of GPC-based compressor trees, area reduction is not as significant as delay, and can be further optimized. In this paper, we propose improved GPC mappings as well as new approaches for GPC cascading and binding for Xilinx FPGAs. With these improvements, we develop an integer linear programming (ILP) method for FPGA synthesis of GPC-based compressor trees that supports cascading and binding between GPCs. Experimental results show that the single-cycle compressor trees produced by the proposed ILP can reduce the average area by 42.40% compared with those generated by existing heuristic method, but are 13.16% slower; the pipelined compressor trees produced by the proposed ILP can reduce the average area by 33.43% at the cost of an average 14.35% decrease in maximum clock frequency compared with those obtained by existing heuristic method.

Recent advances in deep learning, especially deep convolutional neural networks, have led to great performance improvement over semantic segmentation systems. Unfortunately, training deep neural networks (DNNs) requires a humongous amount of labeled data, which is laborious and costly to collect and annotate. Thus, plenty of works have proposed an alternative solution to ease the training set creation by using synthetic data. However, models trained on these kinds of data usually under-perform on real images due to the well-known issue of domain shift. To address it, we propose a generative adversarial network (GAN)-based framework called category-level adversarial adaptation networks (CAA-Nets) for domain adaptation in the context of semantic segmentation. Considering semantic predictions that contain spatial and structure information of images, our idea is to make use of this character by imposing discriminators on the semantic predictions. Different from existing works, the proposed framework utilizes a category-level discriminator in the output space to shrink the gap between real and synthetic images. Similar to reinforcement learning, we take final results as a guide to update parameters in the right direction by use of the output-based discriminator. Moreover, to further enhance the performance, we construct an image-based generator and discriminator pair to distill the feature representations obtained by a DNN. Taking advantage of these modules, our model can achieve competitive performance compared with some existing methods. To showcase the generality and scalability of our approach, we evaluate the proposed method on the Cityscapes dataset by adapting from GTAV and SYNTHIA datasets, where the results demonstrate the effectiveness of our method.

In this paper, we propose a novel image inpainting framework that takes advantage of holistic and structure information of the broken input image. Different from the existing models that complete the broken pictures using the holistic features of the input, our method adopts Patch-generative adversarial networks (GANs) equipped with multi-scale discriminators and edge process function to extract holistic, structured features, and restore the damaged images. After pre-training our Patch-GANs, the proposed network encourages our generator to find the best encoding of the broken input images in the latent space using a combination of a reconstruction loss, an edge loss, and global and local guidance losses. Besides, the reconstruction and the global guidance losses ensure the pixel reliability of the generated images, and the remaining losses guarantee the contents consistency between the local and global parts. The qualitative and quantitative experiments on multiple public datasets show that our approach has the ability to produce more realistic images compared with some existing methods, demonstrating the effectiveness and superiority of our method.

In this paper, we propose a Resetting-Label Network based on Fast Group Loss for Person Re-Identification (RLFGL-ReID). The major challenge of Re-ID lies in how to preserve the similarity of the same person against large variations caused by complex background, different illuminations, and various view angles while discriminating different individuals. To address the above-mentioned problems, we propose the RLFGL-ReID that includes resetting-label (RL) and fast group loss (FGL). Two main contributions of our network are as follows. First, a new method, the resetting-label method, which resets the ID labels, is proposed for the Re-ID network. Resetting the ID labels of each pedestrian is beneficial in maximizing the inter-group distances between each people, achieving better performance on classifying different individuals. Second, a fast group loss, i.e., an advanced version of variance group loss (VGL), is proposed to simplify the training process and accelerate the loss computation. By doing so, the network can eliminate the restriction of inputting the whole group of data when training the network. To confirm the effectiveness of our method, we extensively conduct our method on several widely used person Re-ID benchmark datasets. As the result shows, our method achieves rank@1 accuracy of 98.38% on CUHK03, 95.46% on Market1501, and 91.2% on DukeMTMC-reID, outperforming the state-of-the-art methods and confirming the advantage of our method.

In this Letter, the authors propose a novel attention mechanism combined with a classical generative adversarial network (GAN) model to improve the visual quality of generated samples. This novel attention model is named regional attention GAN. The proposed mechanism can build dependencies between the high‐level representations extracted from attention regions of real images and corresponding feature maps of the generative network. By modelling these dependencies, the generative network can be facilitated to learn feature mapping and fit the distribution of real data. They conduct extensive experiments on widely used datasets CIFAR‐10, STL‐10, and CelebA. The quantitative and qualitative performance improvement over state‐of‐the‐art methods demonstrates the validity of the proposed attention mechanism in improving the quality of generated images.

Multimodal image-to-image translation is a class of vision and graphics problems where the goal is to learn a one-to-many mapping between the source domain and target domain. Given an image in the source domain, the model aims to produce as many diverse results as possible. It is an important and challenging problem in the task of image translation. To this end, recent works utilise Gaussian vectors to produce diverse results but with a small difference. It is because of the special probabilistic nature of Gaussian distribution. In this work, the authors propose linearly distributed latent codes instead of conventional Gaussian vectors, which control the style of generated images. Taking advantage of linear distribution, their model can produce much more diverse results and outperform the state-of-the-art baselines in terms of diversity. Qualitative and quantitative comparisons against baselines demonstrate the effectiveness and superiority of their method.

Recently, digital lock-in amplifiers (DLIAs) have been widely used in scientific research field for measuring weak signals. The method of implementing a DLIA hardware (like field programmable gate array) from the high-level specifications (such as C or C++) is required for software engineers. High-level synthesis (HLS) tool enables the realization of hardware design from high-level specifications automatically. The implementation of the DLIA by HLS is introduced in detail in this paper. The simulation results verify the feasibility of the implementation of the DLIA by HLS.

In this paper, we propose channel and constraint compensation mechanism applied in Generative Adversarial Networks (GANs) to help distribution fitting and improve the visual quality of generated samples. The proposed channel compensation focuses on specific feature-related regions by weighting the channel of conv-layer feature maps, so specific feature modes of data distribution can be compensated and irrelevant features can also be decayed. By combining the Jensen-Shannon (JS) divergence and Wasserstein distance (WD) into a well-designed objective function, the constraint compensation can impose more useful constraints upon the generator to diversify the estimated density in capturing multi-modes. Extensive experiments are conducted on synthetic 2D data and real-world datasets (CIFAR-10, STL-10, CelaBa). The qualitative and quantitative comparisons against baselines demonstrate the effectiveness and superiority of our method.

This paper presents an output-capacitorless adaptively biased (AB) low-dropout regulator (LDO) with stable behavior in 0–25 mA load range. The enhanced multipath nested Miller compensation (EMNMC) with embedded feedforward path and Q-reduction compensation technique are proposed, to achieve high loop gain and wide bandwidth. A symmetrically matched current-voltage mirror (SMCVM) is applied to implement the AB scheme, enabling accurate load current sensing and robust operation of the LDO during full load range. The proposed LDO has been designed in a 0.13-μm CMOS technology. Simulation results verify that the proposed LDO works stably under 0~25 mA loading current and 0~50 pF loading capacitance. It has a maximum current efficiency of 99.17% and a maximum unity-gain frequency (UGF) of 132 MHz. The extension of UGF also magnificently contributes to improvement of power-supply rejection (PSR).

A new circuit topology for pulsed time-of-flight (TOF) laser radar/range-finder receivers to achieve high sensitivity, wide dynamic range and low walk error is presented in this paper. The proposed topology adopts an adaptive structure as pre-amplifier which can work as a trans-impedance amplifier (TIA) or a charge-sensitive amplifier (CSA) depending on the current intensity received by the photodetector. With the help of the adaptive preamplifier, both the dynamic range and the sensitivity can be improved. In addition, a high-accuracy timing discrimination scheme based on converting the incoming unipolar pulse to a bipolar pulse is used to improve the amplitude-dependent timing walk error. Based on the proposed topology, a receiver channel is implemented with a standard 130nm CMOS technology and simulated in Spectre. The CSA and TIA in the receiver channel can share the same core amplifier so that the area and power efficiency are optimized. Simulation results show that the receiver channel achieves an RMS input-referred noise current of 39nA and a timing walk error of ±37ps (±5.6mm in distance) within the wide dynamic range of 1:100000 (100dB), and maximum walk error of ±19mm over the temperature range of −40°C to + 60°C.

The chip-to-board model is important for predicting the electromagnetic emission (EME) of an electronic system and checking whether it fulfills the emission standards at early design cycle. This paper proposes to build the chip-to-board model through directly measuring the chip-to-board transmission ratio with embedded circuit. The embedded circuit does not cause the modification the on-chip power distribution network. Only one signal pin is added to the chip. It preserves the original physical structure of the chip and package in a maximal way. The measurement method is verified through transistor level simulations and is demonstrated with a test chip. Using the measured transmission ratio, one can build a chip-to-board model with improved accuracy, especially in the high frequency range, for the EME simulation of electronic system.

This Letter introduces a weakly supervised method for human action localisation, dubbed Two-stream Action Activation Network (TAAN). In order to generate both action category predictions and temporal location predictions only by video-level annotations, TAAN firstly utilises class response variance to select representative segments as video-level expressions. Then under the supervision of video-level annotations, these expressions are used to train a classification network. Next, temporal class activation maps (T-CAMs) are generated based on the classification response, and finally temporal level predictions are given according to T-CAMs. In addition, the proposed method fuses two-steam predictions in a simple but effective way to further improve the precision of the predictions. Evaluations on THUMOS 2014 dataset show TAAN outperforms state-of-the-art weakly supervised methods.

A new transceiver RFIC for China Geo-Radio Standard is proposed in this paper. It includes all necessary RF blocks to achieve the geo functionality and is implemented in a 0.13-μm RFCMOS technology, with a chip area of 3 mm∗3.4 mm. The power supply voltage for the transceiver is 1.8 V, and the measured maximum output power is 4.4 dBm. The noise figure of the receiver is measured to be 1 dB with external high-performance low-noise amplifier (LNA), and the measured EVM of the transmitter under maximum output power is 2.48%. Measurement results verified that the proposed design is very suitable for the China Geo-Radio Standard.

Distributed arithmetic (DA) algorithm is widely used for finite impulse response (FIR) filter implementation. In the beginning, DA was proposed as sequential DA (SDA), and then was extended to parallel DA (PDA) for higher throughput. This paper presents a novel PDA FIR filter architecture based on 4:2 compressors which can be mapped on Xilinx FPGAs efficiently. On average, our proposed FIR architectures achieve 17.5% reduction in resource usage and 20.7% improvement in performance compared to the state-of-the-art PDA FIR filter. Also, on average, there is 57.9% reduction in resource usage and 23.0% improvement in performance compared to PDA FIR filters generated by Xilinx Coregen.

In this paper, an approach to synthesize compressor trees in High-level Synthesis (HLS) for FPGAs is proposed. Our approach utilizes the bit-level information to improve the compressor tree synthesis. To obtain the bit-level information targeting compressor tree synthesis, a modified bitmask analysis technique based on prior work is proposed. A series of experimental results show that, compared to the existing heuristic, the average reductions of area and delay are 22.96% and 7.05%. The reductions increase to 29.97% and 9.07% respectively, when the carry chains in FPGAs are utilized to implement the compressor trees.

This paper considers the relationship between the static gate delay and the supply voltage of the digital integrated circuit. An empirical equation with physical implication is proposed for calculating the static gate delay. The expression of the gate delay is very simple. It contains only three constants. The calculation includes only one subtraction step, one division step and one addition step. Transistor level simulations are performed to verify the equation. The model matches the experiment results precisely. It is valid for various technologies, gate types, and operation temperatures. The equation can be applied in EDA tools to simulate the timing of the circuit under the PVT variation and the electromagnetic interference.

Effective coordination can dramatically reduce radio interference and avoid packet collisions for multistation wireless local area networks (WLANs). Coordination itself needs consume communication resource and thus competes with data transmission for the limited wireless radio resources. In traditional approaches, control frames and data packets are transmitted in an alternate manner, which brings a great deal of coordination overhead. In this paper, we propose a new communication model where the control frames can be "attacheda to the data transmission. Thus, control messages and data traffic can be transmitted simultaneously and consequently the channel utilization can be improved significantly. We implement the idea in OFDM-based WLANs called hJam, which fully explores the physical layer features of the OFDM modulation method and allows one data packet and a number of control messages to be transmitted together. hJam is implemented on the GNU Radio testbed consisting of eight USRP2 nodes. We also conduct comprehensive simulations and the experimental results show that hJam can improve the WLANs efficiency by up to 200 percent compared with the existing 802.11 family protocols.

Indoor positioning systems have received increasing attention for supporting location-based services in indoor environments. WiFi-based indoor localization has been attractive due to its open access and low cost properties. However, the distance estimation based on received signal strength indicator (RSSI) is easily affected by the temporal and spatial variance due to the multipath effect, which contributes to most of the estimation errors in current systems. In this work, we analyze this effect across the physical layer and account for the undesirable RSSI readings being reported. We explore the frequency diversity of the subcarriers in orthogonal frequency division multiplexing systems and propose a novel approach called FILA, which leverages the channel state information (CSI) to build a propagation model and a fingerprinting system at the receiver. We implement the FILA system on commercial 802.11 NICs, and then evaluate its performance in different typical indoor scenarios. The experimental results show that the accuracy and latency of distance calculation can be significantly enhanced by using CSI. Moreover, FILA can significantly improve the localization accuracy compared with the corresponding RSSI approach.

A novel scaling algorithm is proposed which consists of a bicubic interpolation and an adaptive sharpening filter. The proposed sharpening filter is added to mitigate the blurring effects existing in bicubic interpolation methods. We also verify the scaling quality by taking into account the adaptive technique. Furthermore, we present both the procedures of filtering before and after interpolation in order to reduce the overall computing time. Compared with the previous reported techniques, our method performs better in terms of both quantitative evaluation and visual quality. To achieve the goal of real time and low cost, we describe a pipelined very large scale integration (VLSI) architecture for the implementation of the algorithm. The VLSI architecture of our image scaling processor contains 695 LEs and yields a processing rate of about 165 MHz by using field-programmable gate array (FPGA) technology. Our proposed architecture reduces the amount of gates by 36.8% while achieves an average PSNR increase of 1.5 dB in image quality.

In this paper, we present a method using video codec technology to compress ECG signals. This method exploits both intra-beat and inter-beat correlations of the ECG signals to achieve high compression ratios (CR) and a low percent root mean square difference (PRD). Since ECG signals have both intra-beat and inter-beat redundancies like video signals, which have both intra-frame and inter-frame correlation, video codec technology can be used for ECG compression. In order to do this, some pre-process will be needed. The ECG signals should firstly be segmented and normalized to a sequence of beat cycles with the same length, and then these beat cycles can be treated as picture frames and compressed with video codec technology. We have used records from MIT-BIH arrhythmia database to evaluate our algorithm. Results show that, besides compression efficiently, this algorithm has the advantages of resolution adjustable, random access and flexibility for irregular period and QRS false detection.

This study established a fully automated computer-aided diagnosis (CAD) system for the classification of malignant and benign masses via breast magnetic resonance imaging (BMRI). A breast segmentation method consisting of a preprocessing step to identify the air-breast interfacing boundary and curve fitting for chest wall line (CWL) segmentation was included in the proposed CAD system. The Chan-Vese (CV) model level set (LS) segmentation method was adopted to segment breast mass and demonstrated sufficiently good segmentation performance. The support vector machine (SVM) classifier with ReliefF feature selection was used to merge the extracted morphological and texture features into a classification score. The accuracy, sensitivity, and specificity measurements for the leave-half-case-out resampling method were 92.3%, 98.2%, and 76.2%, respectively. For the leave-one-case-out resampling method, the measurements were 90.0%, 98.7%, and 73.8%, respectively.

The localization of eye centers is a very useful cue for numerous applications like face recognition, facial expression recognition, and the early screening of neurological pathologies. Several methods relying on available light for accurate eye-center localization have been exploited. However, despite the considerable improvements that eye-center localization systems have undergone in recent years, only few of these developments deal with the challenges posed by the profile (non-frontal face). In this paper, we first use the explicit shape regression method to obtain the rough location of the eye centers. Because this method extracts global information from the human face, it is robust against any changes in the eye region. We exploit this robustness and utilize it as a constraint. To locate the eye centers accurately, we employ isophote curvature features, the accuracy of which has been demonstrated in a previous study. By applying these features, we obtain a series of eye-center locations which are candidates for the actual position of the eye-center. Among these locations, the estimated locations which minimize the reconstruction error between the two methods mentioned above are taken as the closest approximation for the eye centers locations. Therefore, we combine explicit shape regression and isophote curvature feature analysis to achieve robustness and accuracy, respectively. In practical experiments, we use BioID and FERET datasets to test our approach to obtaining an accurate eye-center location while retaining robustness against changes in scale and pose. In addition, we apply our method to non-frontal faces to test its robustness and accuracy, which are essential in gaze estimation but have seldom been mentioned in previous works. Through extensive experimentation, we show that the proposed method can achieve a significant improvement in accuracy and robustness over state-of-the-art techniques, with our method ranking second in terms of accuracy. According to our implementation on a PC with a Xeon 2.5Ghz CPU, the frame rate of the eye tracking process can achieve 38 Hz.

The localization of eye centers technique is used in several applications. In this paper, we combine the face shape model to eye center location method to obtaining robustness in eye center localization from a low-resolution image. We make use of shape regression approach and combine it to the eye-center localization using isophote curvature features which has shown its advantage from previous work. In the experiments, we use BioID datasets to test our approach for accurate eye center location and robustness to changes in illumination and pose. We demonstrate that our approach can achieve an improvement of the robustness of eye center localization while the accuracy at eye center localization e≤0.1 and e≤0.15 have achieve 0.918 and 0.977 respectively.

Double-sided DIBR can generate a perfect virtual view image. However, most of the data are the same in the two candidates. It is a waste cost. This paper proposed a simpler and more effective method to reduce the cost called forward and reverse wrapping DIBR (FR-DIBR). The FR-DIBR utilizes a main color image with associated depth map and an assistant color image. First, it wraps the main reference image to the virtual plane, and then reversely wraps the region of the disocclusion in the virtual view to the assistant reference image to get the lost information of the holes. The efficiency is improved by only wrapping one image and the disocclusion region. A depth map inpainting method called BBWF is also proposed. The experimental results show that the FR-DIBR can reduce the computation cost by 33 % and generate a clearer virtual image compared with the double-sided DIBR.

Image scaling is a very important technique and has been widely used in many image and video processing applications. To achieve the goal of low cost and real time, a novel scaling algorithm is proposed which consists of a bilinear interpolation and an adaptive sharpening filter. The proposed sharpening filter is added to perfect the blurring effects existing in traditional bilinear interpolation methods. Simultaneously, we also verify the scaling quality by taking into account the adaptive technology. Compared with the previous bilinear techniques, our method performs better in terms of both quantitative evaluation and visual quality.

Based on the analysis of conventional deinterlacing methods, a new motion adaptive deinterlacing method is proposed. In this paper, we improve three aspects of the conventional motion adaptive method. Firstly, we add a FIR filter to the motion detection part, which not only avoids the wrong detecting of moving objects but also reduces the noises effectively. Secondly, according to the fact, that human eyes are less sensitive to lighter or darker area than gray area, a threshold adaptive method is introduced, which enhanced the subjective quality of deinterlacing. Thirdly, we adopted different interpolation methods for different pixels, which improved the deinterlacing quality of motionless videos. The simulation results show that our proposed deinterlacing method can achieve higher PSNR (peak signal noise ratio) than that of previous studies and can also attain better quality of subjective view.

Bit Plane Coding (BPC) is an important component of the EBCOT block in JPEG2000 encoder. This paper introduced new algorithm for neighboring significance contribution and Run Length Coding (RLC). The paper also presented a fast word level sequential scheme and parallel architecture for bit plane coding. Bit plane coding adopted by EBCOT is mainly divided into three stages: significance generation, coding pass prediction and context formation, which work in parallel and pipelined. The proposed architecture simplifies control procedure and operates at a higher processing speed (up to 164MHz in Xilinx Virtex 4 FPGA).

An high-performance implementation of 2-D lifting-based Discrete Wavelet Transform (DWT) in JPEG2000 applications is designed with low memory and high pipeline architecture. The architecture consists of a row processor module, a column processor module and two memory modules. we present two new row/column processor architecture and memory architecture, one of which includes 7 dual port rams. The For the N*N tile image, only 4N temporal memory are required for 5/3 filter. Symmetric extension is used at the boundaries. Two outputs are generated every cycle. Finally, the proposed architecture was implemented in behavioral verilog-HDL with FPGA (virtex2), the result of which occupied about 400 slices, and was operated in about 120MHz.