arXiv subjects

Feng Gao

Publications and source records attributed to Feng Gao.

At least 235 records · Page 13Linked to original sources

arXiv2022

Change Detection from Synthetic Aperture Radar Images via Graph-Based Knowledge Supplement Network

Synthetic aperture radar (SAR) image change detection is a vital yet challenging task in the field of remote sensing image analysis. Most previous works adopt a self-supervised method which uses pseudo-labeled samples to guide subsequent training and testing. However, deep networks commonly require many high-quality samples for parameter optimization. The noise in pseudo-labels inevitably affects the final change detection performance. To solve the problem, we propose a Graph-based Knowledge Supplement Network (GKSNet). To be more specific, we extract discriminative information from the existing labeled dataset as additional knowledge, to suppress the adverse effects of noisy samples to some extent. Afterwards, we design a graph transfer module to distill contextual information attentively from the labeled dataset to the target dataset, which bridges feature correlation between datasets. To validate the proposed method, we conducted extensive experiments on four SAR datasets, which demonstrated the superiority of the proposed GKSNet as compared to several state-of-the-art baselines. Our codes are available at https://github.com/summitgao/SAR_CD_GKSNet.

cs.CV↗

arXiv2022

Adaptive DropBlock Enhanced Generative Adversarial Networks for Hyperspectral Image Classification

In recent years, hyperspectral image (HSI) classification based on generative adversarial networks (GAN) has achieved great progress. GAN-based classification methods can mitigate the limited training sample dilemma to some extent. However, several studies have pointed out that existing GAN-based HSI classification methods are heavily affected by the imbalanced training data problem. The discriminator in GAN always contradicts itself and tries to associate fake labels to the minority-class samples, and thus impair the classification performance. Another critical issue is the mode collapse in GAN-based methods. The generator is only capable of producing samples within a narrow scope of the data space, which severely hinders the advancement of GAN-based HSI classification methods. In this paper, we proposed an Adaptive DropBlock-enhanced Generative Adversarial Networks (ADGAN) for HSI classification. First, to solve the imbalanced training data problem, we adjust the discriminator to be a single classifier, and it will not contradict itself. Second, an adaptive DropBlock (AdapDrop) is proposed as a regularization method employed in the generator and discriminator to alleviate the mode collapse issue. The AdapDrop generated drop masks with adaptive shapes instead of a fixed size region, and it alleviates the limitations of DropBlock in dealing with ground objects with various shapes. Experimental results on three HSI datasets demonstrated that the proposed ADGAN achieved superior performance over state-of-the-art GAN-based methods. Our codes are available at https://github.com/summitgao/HC_ADGAN

cs.CV↗

arXiv2022

A Thousand Words Are Worth More Than a Picture: Natural Language-Centric Outside-Knowledge Visual Question Answering

Outside-knowledge visual question answering (OK-VQA) requires the agent to comprehend the image, make use of relevant knowledge from the entire web, and digest all the information to answer the question. Most previous works address the problem by first fusing the image and question in the multi-modal space, which is inflexible for further fusion with a vast amount of external knowledge. In this paper, we call for a paradigm shift for the OK-VQA task, which transforms the image into plain text, so that we can enable knowledge passage retrieval, and generative question-answering in the natural language space. This paradigm takes advantage of the sheer volume of gigantic knowledge bases and the richness of pre-trained language models. A Transform-Retrieve-Generate framework (TRiG) framework is proposed, which can be plug-and-played with alternative image-to-text models and textual knowledge bases. Experimental results show that our TRiG framework outperforms all state-of-the-art supervised methods by at least 11.1% absolute margin.

cs.CV↗

arXiv2021

Size-adjustable Ring-shape Photoacoustic Tomography System

Photoacoustic tomography (PAT) combines the advantages of the spectroscopic optical absorption contrast and acoustic resolution with deep penetration, and becomes an important novel biomedical imaging technology for scientific research and clinical diagnosis. In this paper, an imaging size-adjustable PAT system is proposed for clinical applications, which can adapt for different size imaging targets. Comparing with the conventional PAT system setup which with a fixed radius ring shape ultrasound transducer (UT) array, the proposed system, which is based on sectorial ultrasound transducer array (SUTA), is more flexible for different size targets imaging. There are 32 elements for each SUTA, and four SUTAs form a 128-channel UT array for photoacoustic wave detection. The four SUTAs are controlled by four stepper motors, and change the SUTAs distribute positon that adapt for imaging applications. The radius of the proposed system imaging region of interest (ROI) can be adjusted from 50 mm to 90 mm, which is much more flexible than the conventional full ring UT array PAT system. The simulation results generated by the MATLAB k-wave toolbox very well demonstrate the feasibility of the proposed system. To further validate the proposed system for size-adjustable imaging, a vascular mimicking phantom and ex-vivo pork breast with indocyanine green (ICG) injected are imaged to prove its feasibility for clinical applications.

physics.ins-det↗

arXiv2021

Physics-Guided Generative Adversarial Networks for Sea Subsurface Temperature Prediction

Sea subsurface temperature, an essential component of aquatic wildlife, underwater dynamics and heat transfer with the sea surface, is affected by global warming in climate change. Existing research is commonly based on either physics-based numerical models or data based models. Physical modeling and machine learning are traditionally considered as two unrelated fields for the sea subsurface temperature prediction task, with very different scientific paradigms (physics-driven and data-driven). However, we believe both methods are complementary to each other. Physical modeling methods can offer the potential for extrapolation beyond observational conditions, while data-driven methods are flexible in adapting to data and are capable of detecting unexpected patterns. The combination of both approaches is very attractive and offers potential performance improvement. In this paper, we propose a novel framework based on generative adversarial network (GAN) combined with numerical model to predict sea subsurface temperature. First, a GAN-based model is used to learn the simplified physics between the surface temperature and the target subsurface temperature in numerical model. Then, observation data are used to calibrate the GAN-based model parameters to obtain better prediction. We evaluate the proposed framework by predicting daily sea subsurface temperature in the South China sea. Extensive experiments demonstrate the effectiveness of the proposed framework compared to existing state-of-the-art methods.

cs.LG↗

arXiv2021

Synthetic Aperture Radar Image Change Detection via Siamese Adaptive Fusion Network

Synthetic aperture radar (SAR) image change detection is a critical yet challenging task in the field of remote sensing image analysis. The task is non-trivial due to the following challenges: Firstly, intrinsic speckle noise of SAR images inevitably degrades the neural network because of error gradient accumulation. Furthermore, the correlation among various levels or scales of feature maps is difficult to be achieved through summation or concatenation. Toward this end, we proposed a siamese adaptive fusion network for SAR image change detection. To be more specific, two-branch CNN is utilized to extract high-level semantic features of multitemporal SAR images. Besides, an adaptive fusion module is designed to adaptively combine multiscale responses in convolutional layers. Therefore, the complementary information is exploited, and feature learning in change detection is further improved. Moreover, a correlation layer is designed to further explore the correlation between multitemporal images. Thereafter, robust feature representation is utilized for classification through a fully-connected layer with softmax. Experimental results on four real SAR datasets demonstrate that the proposed method exhibits superior performance against several state-of-the-art methods. Our codes are available at https://github.com/summitgao/SAR_CD_SAFNet.

eess.IV↗

arXiv2021

SAR Image Change Detection Based on Multiscale Capsule Network

Traditional change detection methods based on convolutional neural networks (CNNs) face the challenges of speckle noise and deformation sensitivity for synthetic aperture radar images. To mitigate these issues, we proposed a Multiscale Capsule Network (Ms-CapsNet) to extract the discriminative information between the changed and unchanged pixels. On the one hand, the capsule module is employed to exploit the spatial relationship of features. Therefore, equivariant properties can be achieved by aggregating the features from different positions. On the other hand, an adaptive fusion convolution (AFC) module is designed for the proposed Ms-CapsNet. Higher semantic features can be captured for the primary capsules. Feature extracted by the AFC module significantly improves the robustness to speckle noise. The effectiveness of the proposed Ms-CapsNet is verified on three real SAR datasets. The comparison experiments with four state-of-the-art methods demonstrated the efficiency of the proposed method. Our codes are available at https://github.com/summitgao/SAR_CD_MS_CapsNet.

cs.CV↗

arXiv2021

Photoacoustic digital brain: numerical modelling and image reconstruction via deep learning

Photoacoustic tomography (PAT) is a newly developed medical imaging modality, which combines the advantages of pure optical imaging and ultrasound imaging, owning both high optical contrast and deep penetration depth. Very recently, PAT is studied in human brain imaging. Nevertheless, while ultrasound waves are passing through the human skull tissues, the strong acoustic attenuation and aberration will happen, which causes photoacoustic signals' distortion. In this work, we use 10 magnetic resonance angiography (MRA) human brain volumes, and manually segment them to obtain the 2D human brain numerical phantoms for PAT. The numerical phantoms contain six kinds of tissues which are scalp, skull, white matter, gray matter, blood vessel and cerebral cortex. For every numerical phantom, optical properties are assigned to every kind of tissues. Then, Monte-Carlo based optical simulation is deployed to obtain the photoacoustic initial pressure. Then, we made two k-wave simulation cases: one takes inhomogeneous medium and uneven sound velocity into consideration, and the other not. Then we use the sensor data of the former one as the input of U-net, and the sensor data of the latter one as the output of U-net to train the network. We randomly choose 7 human brain PA sinograms as the training dataset and 3 human brain PA sinograms as the testing set. The testing result shows that our method could correct the skull acoustic aberration and obtain the blood vessel distribution inside the human brain satisfactorily.

physics.med-ph↗

arXiv2021

Propionamide (C2H5CONH2): The largest peptide-like molecule in space

Peptide bonds, as the molecular bridges that connect amino acids, are crucial to the formation of proteins. Searches and studies of molecules with embedded peptide-like bonds are thus important for the understanding of protein formation in space. Here we report the first tentative detection of propionamide (C2H5CONH2), the largest peptide-like molecule detected in space toward Sagittarius B2(N1) at a position called N1E that is slightly offset from the continuum peak. A new laboratory measurements of the propionamide spectrum were carried out in the 9-461 GHz, which provide good opportunity to check directly for the transition frequencies of detected interstellar lines of propionamide. Our observing result indicates that propionamide emission comes from the warm, compact cores in Sagittarius B2, in which massive protostellars are forming. The column density of propionamide toward Sgr B2(N1E) was derived to be 1.5\times 10^{16} cm^-2, which is three fifths of that of acetamide, and one nineteenth of that of formamide. This detection suggests that large peptide-like molecules can form and survive during star-forming process and may form more complex molecules in the interstellar medium. The detection of propionamide bodes well for the presence of polypeptides, as well as other complex prebiotic molecules in the interstellar medium.

astro-ph.GA↗

arXiv2021

On-chip erbium-doped lithium niobate microring lasers

Lithium niobate on insulator (LNOI), regarded as an important candidate platform for optical integration due to its excellent nonlinear, electro-optic and other physical properties, has become a research hotspot. Light source, as an essential component for integrated optical system, is urgently needed. In this paper, we reported the realization of 1550-nm band on-chip LNOI microlasers based on erbium-doped LNOI ring cavities with loaded quality factors higher than one million, which were fabricated by using electron beam lithography and inductively coupled plasma reactive ion etching processes. These microlasers demonstrated a low pump threshold of ~20 μW and stable performance under the pump of a 980-nm band continuous laser. Comb-like laser spectra spanning from 1510 nm to 1580 nm were observed in high pump power regime, which lays the foundation of the realization of pulsed laser and frequency combs on rare-earth ion doped LNOI platform. This work has effectively promoted the development of on-chip integrated active LNOI devices.

physics.optics↗

arXiv2021

Searching Water Megamasers By Using Mid-infrared Spectroscopy (I): Possible Mid-infrared Indicators

Water megamasers at 22 GHz with a gas disk configuration in galaxies provide the most precise measurements of supermassive black hole masses, as well as independent constraints on the Hubble constant in the nearby universe. The existence of other maser types, such as jet or outflow masers, represents another tracer for AGN science. However, the detection rate of water megamasers in galaxies is extremely low. Over 40 years, only $\sim$ 160 galaxies are found to harbour maser emission, and $\sim$ 30\% of them show features in their maser emission that indicate a disk-like geometry. Therefore, increasing the detection rate of masers is a crucial task to allow expanding on maser studies. We present a comparison of mid-infrared spectroscopic data between a maser galaxy sample and a Seyfert 2 control sample. We find that maser galaxies show significant peculiarities in their mid-infrared spectra: (1) Maser galaxies tend to present stronger silicate absorption at $τ$ 9.7 $μ$m than the control sample, (2) PAH 11.3 $μ$m emission in maser galaxies is much weaker than in the control sample, (3) spectral indices at 20-30 $μ$m are steeper in maser galaxies than in the control sample and tend to be mid-infrared enhanced population. We conclude that there may be good indicators in mid-infrared and far-infrared which could differentiate maser and non-maser Seyfert 2 galaxies. Upcoming infrared facilities, such as the James Webb Space Telescope, may be able to exploit these and other useful criteria and tracers for water megamaser observations.

astro-ph.GA↗

arXiv2021

Learning the Precise Feature for Cluster Assignment

Clustering is one of the fundamental tasks in computer vision and pattern recognition. Recently, deep clustering methods (algorithms based on deep learning) have attracted wide attention with their impressive performance. Most of these algorithms combine deep unsupervised representation learning and standard clustering together. However, the separation of representation learning and clustering will lead to suboptimal solutions because the two-stage strategy prevents representation learning from adapting to subsequent tasks (e.g., clustering according to specific cues). To overcome this issue, efforts have been made in the dynamic adaption of representation and cluster assignment, whereas current state-of-the-art methods suffer from heuristically constructed objectives with representation and cluster assignment alternatively optimized. To further standardize the clustering problem, we audaciously formulate the objective of clustering as finding a precise feature as the cue for cluster assignment. Based on this, we propose a general-purpose deep clustering framework which radically integrates representation learning and clustering into a single pipeline for the first time. The proposed framework exploits the powerful ability of recently developed generative models for learning intrinsic features, and imposes an entropy minimization on the distribution of the cluster assignment by a dedicated variational algorithm. Experimental results show that the performance of the proposed method is superior, or at least comparable to, the state-of-the-art methods on the handwritten digit recognition, fashion recognition, face recognition and object recognition benchmark datasets.

cs.CV↗

arXiv2021

Integrated LNOI single-mode lasers by Vernier effect

Microcavity lasers based on erbium-doped lithium niobate on insulator (LNOI), which are key devices for LNOI integrated photonics, have attracted much attention recently. In this Letter, we report the realization of a C-band single-mode laser using Vernier effect in two coupled Erbium-doped LNOI microrings with different radii under the pump of a 980-nm continuous laser. The laser, operating stably over a large range of pumping power, has a pump threshold of ~200 μW and a side-mode suppression ratio exceeding 26 dB. The high-performance LNOI single-mode laser will promote the development of lithium niobate integrated photonics.

physics.optics↗

arXiv2021

Towards Efficient Full 8-bit Integer DNN Online Training on Resource-limited Devices without Batch Normalization

Huge computational costs brought by convolution and batch normalization (BN) have caused great challenges for the online training and corresponding applications of deep neural networks (DNNs), especially in resource-limited devices. Existing works only focus on the convolution or BN acceleration and no solution can alleviate both problems with satisfactory performance. Online training has gradually become a trend in resource-limited devices like mobile phones while there is still no complete technical scheme with acceptable model performance, processing speed, and computational cost. In this research, an efficient online-training quantization framework termed EOQ is proposed by combining Fixup initialization and a novel quantization scheme for DNN model compression and acceleration. Based on the proposed framework, we have successfully realized full 8-bit integer network training and removed BN in large-scale DNNs. Especially, weight updates are quantized to 8-bit integers for the first time. Theoretical analyses of EOQ utilizing Fixup initialization for removing BN have been further given using a novel Block Dynamical Isometry theory with weaker assumptions. Benefiting from rational quantization strategies and the absence of BN, the full 8-bit networks based on EOQ can achieve state-of-the-art accuracy and immense advantages in computational cost and processing speed. What is more, the design of deep learning chips can be profoundly simplified for the absence of unfriendly square root operations in BN. Beyond this, EOQ has been evidenced to be more advantageous in small-batch online training with fewer batch samples. In summary, the EOQ framework is specially designed for reducing the high cost of convolution and BN in network training, demonstrating a broad application prospect of online training in resource-limited devices.

cs.LG↗

arXiv2021

Photoacoustic-monitored laser treatment for tattoo removal: a feasibility study

Skin blemishes and diseases have attracted increasing research interest in recent decades, due to their growing frequency of occurrence and the severity of related diseases. Various laser treatment approaches have been introduced for the alleviation and removal of skin pigmentation. The treatments' effects highly depend on the experience and prognosis of the relevant operators. But, the operation process lacks real-time feedback, which may directly reflect the extent of the treatment. In this manuscript, we report a photoacoustic-guided laser treatment method with a feasibility study, specifically for laser treatment targeting the tattoo's removal. The results well validated the feasibility of the proposed method through the experiments on phantoms and ex vivo pig skin samples.

eess.SP↗

arXiv2021

Deep Learning Enabled Real-Time Photoacoustic Tomography System via Single Data Acquisition Channel

Photoacoustic computed tomography (PACT) combines the optical contrast of optical imaging and the penetrability of sonography. In this work, we develop a novel PACT system to provide real-time imaging, which is achieved by a 120-elements ultrasound array only using a single data acquisition (DAQ) channel. To reduce the channel number of DAQ, we superimpose 30 nearby channels' signals together in the analog domain, and shrinking to 4 channels of data (120/30=4). Furthermore, a four-to-one delay-line module is designed to combine these four channels' data into one channel before entering the single-channel DAQ, followed by decoupling the signals after data acquisition. To reconstruct the image from four superimposed 30-channels'PA signals, we train a dedicated deep learning model to reconstruct the final PA image. In this paper, we present the preliminary results of phantom and in-vivo experiments, which manifests its robust real-time imaging performance. The significance of this novel PACT system is that it dramatically reduces the cost of multi-channel DAQ module (from 120 channels to 1 channel), paving the way to a portable, low-cost and real-time PACT system.

eess.IV↗

arXiv2021

Capability Iteration Network for Robot Path Planning

Path planning is an important topic in robotics. Recently, value iteration based deep learning models have achieved good performance such as Value Iteration Network(VIN). However, previous methods suffer from slow convergence and low accuracy on large maps, hence restricted in path planning for agents with complex kinematics such as legged robots. Therefore, we propose a new value iteration based path planning method called Capability Iteration Network(CIN). CIN utilizes sparse reward maps and encodes the capability of the agent with state-action transition probability, rather than a convolution kernel in previous models. Furthermore, two training methods including end-to-end training and training capability module alone are proposed, both of which speed up convergence greatly. Several path planning experiments in various scenarios, including on 2D, 3D grid world and real robots with different map sizes are conducted. The results demonstrate that CIN has higher accuracy, faster convergence, and lower sensitivity to random seed compared to previous VI-based models, hence more applicable for real robot path planning.

cs.RO↗

arXiv2021

Change Detection in Synthetic Aperture Radar Images Using a Dual-Domain Network

Change detection from synthetic aperture radar (SAR) imagery is a critical yet challenging task. Existing methods mainly focus on feature extraction in spatial domain, and little attention has been paid to frequency domain. Furthermore, in patch-wise feature analysis, some noisy features in the marginal region may be introduced. To tackle the above two challenges, we propose a Dual-Domain Network. Specifically, we take features from the discrete cosine transform domain into consideration and the reshaped DCT coefficients are integrated into the proposed model as the frequency domain branch. Feature representations from both frequency and spatial domain are exploited to alleviate the speckle noise. In addition, we further propose a multi-region convolution module, which emphasizes the central region of each patch. The contextual information and central region features are modeled adaptively. The experimental results on three SAR datasets demonstrate the effectiveness of the proposed model. Our codes are available at https://github.com/summitgao/SAR_CD_DDNet.

cs.CV↗