RGB Mode Research Articles

Hydride generation smartphone readable colorimetric system for speciation analysis of inorganic antimony without chromatographic separation

Antimony (Sb) was considered as a toxic metalloid element in environmental water system, the concentration of antimony in water sample was required to not be higher than 0.02 μg/mL. Miniaturized/portable analytical systems/devices were needed for on-site analysis of antimony. Herein, hydride generation was combined with a smartphone readable colorimetric method and used for the speciation analysis of inorganic antimony without chromatographic separation. Sb(III) was analyzed when masking reagent was used for Sb(V), total antimony concentration was obtained when Sb(V) was educed to Sb(III) before analysis. The conversion of Sb(Ⅲ) to SbH3 was achieved through hydride generation (HG) reaction, a redox reaction between SbH3 and HAuCl4 occurred with a result of in-situ generation of gold nanoparticles (Au NPs) on test paper, changing color from pale yellow to black and allowing for colorimetric analysis with naked-eye and smartphone RGB mode. Detectable limits of 0.02 μg/mL and 0.01 μg/mL can be recognized by naked-eye and smartphone mode, respectively. This miniaturized and easy-used smartphone readable analytical system was further used for on-site analysis of inorganic antimony in real samples, recoveries of 95–105 % were obtained.

Three-dimensional dynamic gesture recognition method based on convolutional neural network

With the rapid advancement of virtual reality, dynamic gesture recognition technology has become an indispensable and critical technique for users to achieve human–computer interaction in virtual environments. The recognition of dynamic gestures is a challenging task due to the high degree of freedom and the influence of individual differences and the change of gesture space. To solve the problem of low recognition accuracy of existing networks, an improved dynamic gesture recognition algorithm based on ResNeXt architecture is proposed. The algorithm employs three-dimensional convolution techniques to effectively capture the spatiotemporal features intrinsic to dynamic gestures. Additionally, to enhance the model’s focus and improve its accuracy in identifying dynamic gestures, a lightweight convolutional attention mechanism is introduced. This mechanism not only augments the model’s precision but also facilitates faster convergence during the training phase. In order to further optimize the performance of the model, a deep attention submodule is added to the convolutional attention mechanism module to strengthen the network’s capability in temporal feature extraction. Empirical evaluations on EgoGesture and NvGesture datasets show that the accuracy of the proposed model in dynamic gesture recognition reaches 95.03% and 86.21%, respectively. When operating in RGB mode, the accuracy reached 93.49% and 80.22%, respectively. These results underscore the effectiveness of the proposed algorithm in recognizing dynamic gestures with high accuracy, showcasing its potential for applications in advanced human–computer interaction systems.

Smartphone-assisted colorimetric sensor arrays based on nanozymes for high throughput identification of heavy metal ions in salmon

The rapid, precise, and high-throughput identification of multiple heavy metals ions holds immense importance in ensuring food safety and promoting public health. This study presents a novel smartphone-assisted colorimetric sensor array for the rapid and precise detection of multiple heavy metals ions. The sensor array is based on three signal recognition elements (AuPt@Fe-N-C, AuPt@N-C, and Fe-N-C) and the presence of different heavy metal ions affects the nanozymes-chromogenic substrate (TMB) catalytic color production, enabling the differentiation and quantification of various heavy metal ions. Combined with a smartphone-based RGB mode, the colorimetric sensor array can successfully identify five different heavy metal ions (Hg2+, Pb2+, Co2+, Cr6+, and Fe3+) as low as 0.5 μM and different ratios of binary and ternary mixed heavy metal ions in just 5 min. The sensor array successfully tested seawater and salmon samples with a total heavy metal content of 10 μM in the South China Sea (Haikou and Wenchang). Overall, this study highlights the potential of smartphone-assisted colorimetric sensor arrays for the rapid and precise detection of multiple heavy metal ions, which could significantly contribute to food safety and public health monitoring.

A dual-signal triple-readout optical sensing platform for α-glucosidase and β-glucosidase activity monitoring and inhibitor screening based on luminescent covalent organic framework

BackgroundAs promising biomarkers of diabetes, α-glucosidase (α-Glu) and β-glucosidase (β-Glu) play a crucial role in the diagnosis and management of diseases. However, there is a scarcity of techniques available for simultaneously and sensitively detecting both enzymes. What's more, most of the approaches for detecting α-Glu and β-Glu rely on a single-mode readout, which can be affected by multiple factors leading to inaccurate results. Hence, the simultaneous detection of the activity levels of both enzymes in a single sample utilizing multiple-readout sensing approaches is highly attractive. ResultsIn this work, we constructed a facile sensing platform for the simultaneous determination of α-Glu and β-Glu by utilizing a luminescent covalent organic framework (COF) as a fluorescent indicator. The enzymatic hydrolysis product common to both enzymes, p-nitrophenol (PNP), was found to affect the fluorometric signal through an inner filter effect on COF, enhance the colorimetric response by intensifying the absorption peak at 400 nm, and induce changes in RGB values when analyzed using a smartphone-based color recognition application. By combining fluorometric/colorimetric measurements with smartphone-assisted RGB mode, we achieved sensitive and accurate quantification of α-Glu and β-Glu. The limits of detection for α-Glu were determined to be 0.8, 1.22, and 1.85 U/L, respectively. Similarly, the limits of detection for β-Glu were 0.16, 0.42, and 0.53 U/L, respectively. SignificanceApplication of the proposed sensing platform to clinical serum samples revealed significant differences in the two enzymes between healthy people and diabetic patients. Additionally, the proposed sensing method was successfully applied for the screening of α-Glu inhibitors and β-Glu inhibitors, demonstrating its viability and prospective applications in the clinical management of diabetes as well as the discovery of antidiabetic medications.

Performance Optimization for Multicolor Electrochromic Devices: Morphology and Composition Regulation of Inorganic Electrochromic Materials and Selectivity of Ions

AbstractIn recent years, multicolor electrochromic devices (ECDs) have attracted much attention due to their application potential for rich color similar to RGB mode. Although organic electrochromic (EC) materials are widely used in multicolor electrochromism because of their multicolor properties, remarkable weatherability remains challenging. Inorganic EC materials especially transition metal oxide are particularly attractive for excellent weatherability in ECD fabrication. Moreover, by selection of electrolytes, better ionic transfer efficiency can be rendered in the EC process. Although the morphology regulation of the EC film and the selection of electrolytes in the ion transmission layer play important roles in optimizing EC performance and have been extensively studied, a systematic and comprehensive review is lacking. In this review, the morphology regulation of the EC film and selection rules of electrolytes in the multicolor ECDs is focused, and the advantages and disadvantages of existing ways are discussed. The current application of multicolor ECDs is also outlined and it is hoped that this review can contribute some inspiration to designing and improving multicolor ECDs.

Portrait Semantic Segmentation Method Based on Dual Modal Information Complementarity

Semantic segmentation of human images is a research hotspot in the field of computer vision. At present, the semantic segmentation models based on U-net generally lack the ability to capture the spatial information of images. At the same time, semantic incompatibility exists because the feature maps of encoder and decoder are directly connected in the skip connection stage. In addition, in low light scenes such as at night, it is easy for false segmentation and segmentation accuracy to appear. To solve the above problems, a portrait semantic segmentation method based on dual-modal information complementarity is proposed. The encoder adopts a double branch structure, and uses a SK-ASSP module that can adaptively adjust the convolution weights of different receptor fields to extract features in RGB and gray image modes respectively, and carries out cross-modal information complementarity and feature fusion. A hybrid attention mechanism is used in the jump connection phase to capture both the channel and coordinate context information of the image. Experiments on human matting dataset show that the PA and MIoU coefficients of this algorithm model reach 96.58% and 94.48% respectively, which is better than U-net benchmark model and other mainstream semantic segmentation models.

Nonmetal catalyst boosting amplification of both colorimetric and electrochemical signal for multi-mode nitrite sensing

Recently, nanozymes as an outstanding alternative to natural enzymes has attracted wide attention because of its high stability performance. In this study, PNC nonmetal nanozymes with high oxidase-like activity was synthesized can specifically catalyze colorless 3,3,5,5-tetramethyl-benzidine(TMB) to form blue oxidized TMB (TMBox). In the presence of nitrite, it further oxidizes TMBox to obtain yellow derivative products attributed to nitrite inducing diazotization reaction in TMBox. Based on this principle, a colorimetric and electrochemical sensing system was developed, and the ultra-sensitive multi-mode detection of nitrite was realized by combining RGB mode of smart phone, UV–Vis spectrum and electrochemical method. Compared with single signal detection, the multi-mode sensing system can realize self-validation to achieve more reliable detection results. What's more, the developed multi-mode sensing could quickly and sensitively detect nitrite in real sample, especially RGB mode of smart phone meeting the equipment limited areas, suggesting a broad application prospects in food safety.

Sensitive visual detection of norfloxacin in water by smartphone assisted colorimetric method based on peroxidase-like active cobalt-doped Fe3O4 nanozyme

Norfloxacin is widely used owing to its strong bactericidal effect on Gram-negative bacteria. However, the residual norfloxacin in the environment can be biomagnified via food chain and may damage the human liver and delay the bone development of minors. Present work described a reliable and sensitive smartphone colorimetric sensing system based on cobalt-doped Fe3O4 magnetic nanoparticles (Co-Fe3O4 MNPs) for the visual detection of norfloxacin. Compared with Fe3O4, Co-Fe3O4 MNPs earned more remarkably peroxidase-like activity and TMB (colorless) was rapidly oxidized to oxTMB (blue) with the presence of H2O2. Interestingly, the addition of low concentration of norfloxacin can accelerate the color reaction process of TMB, and blue deepening of the solution can be observed with the naked eye. However, after adding high concentration of norfloxacin, the activity of nanozyme was inhibited, resulting in the gradual fading of the solution. Based on this principle, a colorimetric sensor integrated with smartphone RGB mode was established. The visual sensor exhibited good linearity for norfloxacin monitoring in the range of 0.13-2.51 µmol/L and 17.5-100 µmol/L. The limit of visual detection was 0.08 µmol/L. In the actual water sample analysis, the spiked recoveries of norfloxacin were over the range of 95.7%-104.7 %. These results demonstrated that the visual sensor was a convenient and fast method for the efficient and accurate detection of norfloxacin in water, which may have broad application prospect.

Sub-Pixel Convolutional Neural Network for Image Super-Resolution Reconstruction

Image super-resolution (SR) reconstruction technology can improve the quality of low-resolution (LR) images. There are many available deep learning networks different from traditional machine learning algorithms. However, these networks are usually prone to poor performance on complex computation, vanishing gradients, and loss of useful information. In this work, we propose a sub-pixel convolutional neural network (SPCNN) for image SR reconstruction. First, to reduce the strong correlation, the RGB mode was translated into YCbCr mode, and the Y channel data was chosen as the input LR image. Meanwhile, the LR image was chosen as the network input to reduce computation instead of the interpolation reconstructed image as used in the super-resolution convolutional neural network (SRCNN). Then, two convolution layers were built to obtain more features, and four non-linear mapping layers were used to achieve different level features. Furthermore, the residual network was introduced to transfer the feature information from the lower layer to the higher layer to avoid the gradient explosion or vanishing gradient phenomenon. Finally, the sub-pixel convolution layer based on up-sampling was designed to reduce the reconstruction time. Experiments on three different data sets proved that the proposed SPCNN performs superiorly to the Bicubic, sparsity constraint super-resolution (SCSR), anchored neighborhood regression (ANR), and SRCNN methods on reconstruction precision and time consumption.

In-line gas purging exploiting a novel converging flow-on paper optosensing device for the determination of total ammonia-nitrogen in water samples

A novel, inexpensive, and sensitive microanalytical system for the determination of total ammonia-nitrogen (ammonia-N) in water was developed. The system was designed to detect ammonia using an in-line purging system coupled with a newly designed gas converging flow cell incorporating in-line paper-based colorimetric detection. The gas converging flow cell served as a gas chamber, a paper-based device holder, and a color detection cell. The paper-based device was impregnated with butterfly pea extract (a natural acid-base indicator), the color of which changed on the basis of the generated ammonia gas. Analytical parameters such as concentration of natural indicator and purging time were studied and optimized. The color image could be captured by an integrated endoscope camera or smartphone and analyzed by ImageJ software in RGB mode. Under optimal conditions, two linear dynamic ranges were observed. By using 1:2 (v/v) diluted 2% (w/v) butterfly pea extract in Milli-Q water, a low linear concentration range of 0.05–0.9 mg L−1 with a limit of detection (LOD) of 0.02 mg L−1 and limit of quantification (LOQ) of 0.05 mg L−1 was achieved. By using a non-diluted 2% (w/v) butterfly pea extract, a high linear concentration range of 0.5–4 mg L−1 with an LOD of 0.1 mg L−1 and LOQ of 0.3 mg L−1 was obtained. Finally, the optimized method was leveraged to determine ammonia-N in various water samples from shrimp farming.

Data Security Using Color Image Based on Beaufort Cipher, Column Transposition and Least Significant Bit (LSB)

One of cryptography algorithm which used is beaufort cipher. Beaufort cipher has simple encryption procedure, but this algorithm has good enough endurance to attack. Unauthorized people cannot break up decrypt without know matrix key used. This algorithm used to encrypt data in the form of text called plaintext. The result of this algorithm is string called ciphertext which difficult to understood that can causing suspicious by other people. Beaufort cipher encryption tested with avalanche effect algorithm with modified one, two, three and all key matrix which resulting maximum 31.25% with all key modification so another algorithm is needed to get more secure. Least Significant Bit (LSB) used to insert ciphertext created to form of image. LSB chosen because easy to use and simple, just alter one of last bit image with bit from message. LSB tested with RGB, CMYK, CMY and YUV color modes inserted 6142 characters resulting highest PSNR value 51.2546 on YUV color mode. Applying steganography technique has much advantage in imperceptibility, for example the image product very similar with original cover image so the difference can not differentiate image with human eye vision. Image that tested as much ten images, that consist of five 512 x 512 and five 16 x 16 image. While string message that used is 240, 480 and 960 character to test 512 x 512 image and 24, 48 and 88 character to test 16 x 16 image. The result of experiment measured with Mean Square Error (MSE) and Peak Signal Ratio (PSNR) which has minimum PSNR 51.2907 dB it means stego image that produced hood enough. Computation time calculation using tic toc in matlab resulting fastest value 0.041636 to encrypt 2000 character and the longest time is 4.10699 second to encrypt 6000 character and inserting to image. Amount of character and amount of multi algorithm can affecting computation time calculation.

Characterization of a new radiochromic film (LD-V1) using mammographic beam qualities

PurposeRadiochromic film (RCF) is a detector that can obtain a two-dimensional dose distribution with high resolution; it is widely used in medical and industrial fields. Several types of RCFs exist based on their application. The type of RCF mainly used for mammography dose assessment has been discontinued; however, a new type of RCF (LD-V1) has been distributed as a successor. Since the medical use of LD-V1 has rarely been studied, we investigated the response characteristics of LD-V1 in mammography. MethodsMeasurements were performed using Mo/Mo and Rh/Ag on a Senographe Pristina mammography device (GE, Fairfield, CT, USA). The reference air kerma was measured using a parallel-plate ionization chamber (PPIC) (C-MA, Applied Engineering Inc, Tokyo, Japan). Pieces of LD-V1 film model were irradiated at the same position where the reference air kerma in air was measured by the PPIC. Irradiation was performed using the time scale method based on the load on the equipment. Two methods of irradiation were considered: placing the detector in air and on the phantom. The LD-V1 was scanned five times at 72 dpi in RGB (48 bit) mode using a flatbed scanner (ES-G11000, Seiko Epson Corp, Nagano, Japan) 24 h following irradiation. The response ratio of the reference air kerma and the air kerma obtained from the LD-V1 were compared and examined for each beam quality and air kerma range. Results and discussionWhen the beam quality was altered, the response ratio varied from 0.8 to 1.2 with respect to the measurement value of the PPIC; however, some outliers were observed. Response ratios were highly variable in the low-dose range; however, as the air kerma increased, the ratios approached 1. Thus, LD-V1 does not need calibration for each beam quality used in mammography. LD-V1 enables air kerma evaluation by creating air kerma response curves under certain X-ray conditions used in mammography. ConclusionWe suggest that the dose range be limited to 12 mGy or more to keep the response variation with beam qualities below ±20%. If further measurement is required for reducing the response variation, the dose range should be shifted to a higher dose range.

Dual-Signal Triple-Mode Optical Sensing Platform for Assisting in the Diagnosis of Kidney Disorders.

As known biomarkers of kidney diseases, N-acetyl-β-d-glucosaminidase (NAG) and β-galactosidase (β-GAL) are of great importance for the diagnosis and treatment of diseases. The feasibility of using multiplex sensing methods to simultaneously report the outcome of the two enzymes in the same sample is even more alluring. Herein, we establish a simple sensing platform for the concurrent detection of NAG and β-GAL using silicon nanoparticles (SiNPs) as a fluorescent indicator synthesized by a one-pot hydrothermal route. p-Nitrophenol (PNP), as a common enzymatic hydrolysis product of the two enzymes, led to the attenuation of fluorometric signal caused by the inner filter effect on SiNPs, the enhancement of colorimetric signal due to the increase of intensity of the characteristic absorption peak at around 400 nm with increasing reaction time, and the changes of RGB values of images obtained through a color recognition application on a smartphone. The fluorometric/colorimetric approach combined with the smartphone-assisted RGB mode was able to detect NAG and β-GAL with good linear response. Applying this optical sensing platform to clinical urine samples, we found that the two indicators in healthy individuals and patients (glomerulonephritis) with kidney diseases were significantly different. By expanding to other renal lesion-related specimens, this tool may show great potentials in clinical diagnosis and visual inspection.

GaitReload: A Reloading Framework for Defending Against On-Manifold Adversarial Gait Sequences

Recent on-manifold adversarial attacks can mislead gait recognition by generating adversarial walking postures (AWP) with image generation techniques. However, existing defense methods only eliminate adversarial perturbations on each frame isolatedly but ignore the temporal correlation of gait sequence, which leads to vulnerability of robust gait recognition. In this paper, we propose GaitReload, a post-processing adversarial defense method to defend against AWP for the gait recognition model with sequenced inputs. First, GaitReload utilizes sequenced entity recognition (SER) module to detect the adversarial frames by the temporal constraints of gait sequence. Then, we apply bayesian uncertainty filtering-based (BUF-based) gait interpolation to reform adversarial gait examples. After that, we reload the reformed gait sequence and rectify the recognition results with the guidance of reloading strategy. Specifically, SER has a bi-directional frame difference attention and a temporal feature aggregation to boost the detection performance. For training SER, we apply hidden posture selective attack (HPSA) to generate training samples. The extensive experimental results on CASIA-A, CASIA-B, and OU-ISIR demonstrate that GaitReload can defend against adversarial gait by large margins in both RGB and silhouette modes.

A colorimetric smartphone-based sensor for on-site AA detection in tropical fruits using Fe-P/N[sbnd]C single-atom nanoenzyme

Ascorbic acid is one of the important vitamins to maintain human life activities and plays an irreplaceable role in regulating human redox metabolism. Fresh fruit can provide plenty of AA to maintain human metabolic balance. Thus, it is great significant to develop a rapid and convenient method for detection of AA to evaluate the freshness and nutritional quality of fruits. In this work, Fe single-atom nanoenzyme (Fe-SAN) based colorimetric sensor assisted with smartphone was designed for rapid and on-site AA detection in tropical fruits. Firstly, Fe-SAN with high oxidase-mimicking activity was synthesized by using green tea leaves as sources of carbon and nitrogen and NaH2PO2 as P source to obtain Fe-P/NC SAN, in which P was used to reconstruct the distribution of electronic to enhance the oxidase-mimicking activity of Fe-SAN. Besides, the as-synthesized Fe-P/NC SAN with remarkable oxidase-like activities could oxidize 3,3́,5,5́-tetramethylbenzidine (TMB) to blue colored oxidized TMB. AA could inhibit the oxidation of TMB, leading to blue fading. Based on the above principle, colorimetric sensor integrated with smartphone RGB mode was fabricated and exhibited a good linear detection range (0.5–100 μM) and low detection limit of 0.315 μM for AA detection under optimal conditions.More importantly, the developed sensor could rapidly and accurately detect AA in real sample, such as pineapple, wax apple and mango. Therefore, this research provides a new cost-effective method for the efficient and exact detection of AA in tropical fruit, which has a broad application prospect.

SmartAct: Energy Efficient and Real-Time Hand-to-Mouth Gesture Detection Using Wearable RGB-T.

Researchers have been leveraging wearable cameras to both visually confirm and automatically detect individuals' eating habits. However, energy-intensive tasks such as continuously collecting and storing RGB images in memory, or running algorithms in real-time to automate detection of eating, greatly impacts battery life. Since eating moments are spread sparsely throughout the day, battery life can be mitigated by recording and processing data only when there is a high likelihood of eating. We present a framework comprising a golf-ball sized wearable device using a low-powered thermal sensor array and real-time activation algorithm that activates high-energy tasks when a hand-to-mouth gesture is confirmed by the thermal sensor array. The high-energy tasks tested are turning on the RGB camera (Trigger RGB mode) and running inference on an on-device machine learning model (Trigger ML mode). Our experimental setup involved the design of a wearable camera, 6 participants collecting 18 hours of data with and without eating, the implementation of a feeding gesture detection algorithm on-device, and measures of power saving using our activation method. Our activation algorithm demonstrates an average of at-least 31.5% increase in battery life time, with minimal drop of recall (5%) and without impacting the accuracy of detecting eating (a slight 4.1% increase in F1-Score).

Rational construction of deep-red fluorescent probe for rapid detection of HClO and its application in bioimaging and paper-based sensing

Hypochlorous acid (HClO), the core bactericidal substance of the human immune system, plays a vital role in many physiological and pathological processes in the human body. In this work, we designed and synthesized a novel deep-red fluorescent probe TCF-ClO for the determination of hypochlorous acid through theoretical analysis. The results showed that probe TCF-ClO exhibited excellent characteristics of long-wavelength emission (635nm), fast response (< 1min), and low detection limit (24nM). In addition, it had been successfully used for imaging of HClO in living HeLa cells. More importantly, the TCF-ClO composited paper-based sensing material was successfully constructed. The RGB/gray value was obtained from a mobile phone and computer, which could achieve the quantitative detection of HClO, with a linear detection range of 0-50μM and a detection limit of 1.09μM (RGB mode)/3.38μM (gray mode). The function of the paper-based sensor extended from qualitative to quantitative detection of HClO, and it is expected to become a portable device widely used in the environmental area.

Practical Secret Image Sharing Based on the Chinese Remainder Theorem

Compared with Shamir’s original secret image sharing (SIS), the Chinese-remainder-theorem-based SIS (CRTSIS) generally has the advantages of a lower computation complexity, lossless recovery and no auxiliary encryption. However, general CRTSIS is neither perfect nor ideal, resulting in a narrower range of share pixels than that of secret pixels. In this paper, we propose a practical and lossless CRTSIS based on Asmuth and Bloom’s threshold algorithm. To adapt the original scheme for grayscale images, our scheme shares the high seven bits of each pixel and utilizes the least significant bit (LSB) matching technique to embed the LSBs into the random integer that is generated in the sharing phase. The chosen moduli are all greater than 255 and the share pixels are in the range of [0, 255] by a screening operation. The generated share pixel values are evenly distributed in the range of [0, 255] and the selection of (k,n) threshold is much more flexible, which significantly improves the practicality of CRTSIS. Since color images in RGB mode are made up of three channels, it is easy to extend the scheme to color images. Theoretical analysis and experiments are given to validate the effectiveness of the proposed scheme.

A Fast Text-to-Image Encryption-Decryption Algorithm for Secure Network Communication

Data security is the science of protecting data in information technology, including authentication, data encryption, data decryption, data recovery, and user protection. To protect data from unauthorized disclosure and modification, a secure algorithm should be used. Many techniques have been proposed to encrypt text to an image. Most past studies used RGB layers to encrypt text to an image. In this paper, a Text-to-Image Encryption-Decryption (TTIED) algorithm based on Cyan, Magenta, Yellow, Key/Black (CMYK) mode is proposed to improve security, capacity, and processing time. The results show that the capacity increased from one to four times compared to RGB mode. Security was also improved due to a decrease in the probability of an adversary discovering keys. The processing time ranged between 0.001 ms (668 characters) and 31 s (25 million characters), depending on the length of the text. The compression rate for the encrypted file was decreased compared to WinRAR. In this study, Arabic and English texts were encrypted and decrypted.

An End-to-end Heterogeneous Restraint Network for RGB-D Cross-modal Person Re-identification

The RGB-D cross-modal person re-identification (re-id) task aims to identify the person of interest across the RGB and depth image modes. The tremendous discrepancy between these two modalities makes this task difficult to tackle. Few researchers pay attention to this task, and the deep networks of existing methods still cannot be trained in an end-to-end manner. Therefore, this article proposes an end-to-end module for RGB-D cross-modal person re-id. This network introduces a cross-modal relational branch to narrow the gaps between two heterogeneous images. It models the abundant correlations between any cross-modal sample pairs, which are constrained by heterogeneous interactive learning. The proposed network also exploits a dual-modal local branch, which aims to capture the common spatial contexts in two modalities. This branch adopts shared attentive pooling and mutual contextual graph networks to extract the spatial attention within each local region and the spatial relations between distinct local parts, respectively. Experimental results on two public benchmark datasets, that is, the BIWI and RobotPKU datasets, demonstrate that our method is superior to the state-of-the-art. In addition, we perform thorough experiments to prove the effectiveness of each component in the proposed method.