Multilayer Perceptron-Based Error Compensation for Automatic On-the-Fly Camera Orientation Estimation Using a Single Vanishing Point from Road Lane.

This study introduces a multilayer perceptron (MLP) error compensation method for real-time camera orientation estimation, leveraging a single vanishing point and road lane lines within a steady-state framework. The research emphasizes cameras with a roll angle of 0°, predominant in autonomous vehicle contexts. The methodology estimates pitch and yaw angles using a single image and integrates two Kalman filter models with inputs from image points (u, v) and derived angles (pitch, yaw). Performance metrics, including avgE, minE, maxE, ssE, and Stdev, were utilized, testing the system in both simulator and real-vehicle environments. The outcomes indicate that our method notably enhances the accuracy of camera orientation estimations, consistently outpacing competing techniques across varied scenarios. This potency of the method is evident in its adaptability and precision, holding promise for advanced vehicle systems and real-world applications.

Implementation of Field-Programmable Gate Array Platform for Object Classification Tasks Using Spike-Based Backpropagated Deep Convolutional Spiking Neural Networks.

This paper investigates the performance of deep convolutional spiking neural networks (DCSNNs) trained using spike-based backpropagation techniques. Specifically, the study examined temporal spike sequence learning via backpropagation (TSSL-BP) and surrogate gradient descent via backpropagation (SGD-BP) as effective techniques for training DCSNNs on the field programmable gate array (FPGA) platform for object classification tasks. The primary objective of this experimental study was twofold: (i) to determine the most effective backpropagation technique, TSSL-BP or SGD-BP, for deeper spiking neural networks (SNNs) with convolution filters across various datasets; and (ii) to assess the feasibility of deploying DCSNNs trained using backpropagation techniques on low-power FPGA for inference, considering potential configuration adjustments and power requirements. The aforementioned objectives will assist in informing researchers and companies in this field regarding the limitations and unique perspectives of deploying DCSNNs on low-power FPGA devices. The study contributions have three main aspects: (i) the design of a low-power FPGA board featuring a deployable DCSNN chip suitable for object classification tasks; (ii) the inference of TSSL-BP and SGD-BP models with novel network architectures on the FPGA board for object classification tasks; and (iii) a comparative evaluation of the selected spike-based backpropagation techniques and the object classification performance of DCSNNs across multiple metrics using both public (MNIST, CIFAR10, KITTI) and private (INHA_ADAS, INHA_KLP) datasets.

Hydrothermal synthesis of CuO@MnO2 on nitrogen-doped multiwalled carbon nanotube composite electrodes for supercapacitor applications.

Nitrogen-doped multiwalled carbon nanotubes (N-MWCNTs) have been used to fabricate nanostructured materials for various energy devices, such as supercapacitors, sensors, batteries, and electrocatalysts. Nitrogen-doped carbon-based electrodes have been widely used to improve supercapacitor applications via various chemical approaches. Based on previous studies, CuO@MnO2 and CuO@MnO2/N-MWCNT composites were synthesized using a sonication-supported hydrothermal reaction process to evaluate their supercapacitor properties. The structural and morphological properties of the synthesized composite materials were characterized via Raman spectroscopy, XRD, SEM, and SEM-EDX, and the morphological properties of the composite materials were confirmed by the nanostructured composite at the nanometer scale. The CuO@MnO2 and CuO@MnO2/N-MWCNT composite electrodes were fabricated in a three-electrode configuration, and electrochemical analysis was performed via CV, GCD, and EIS. The composite electrodes exhibited the specific capacitance of ~ 184 F g-1 at 0.5 A g-1 in the presence of a 5 M KOH electrolyte for the three-electrode supercapacitor application. Furthermore, it exhibited significantly improved specific capacitances and excellent cycling stability up to 5000 GCD cycles, with a 98.5% capacity retention.

Exploring Optimized Spiking Neural Network Architectures for Classification Tasks on Embedded Platforms.

In recent times, the usage of modern neuromorphic hardware for brain-inspired SNNs has grown exponentially. In the context of sparse input data, they are undertaking low power consumption for event-based neuromorphic hardware, specifically in the deeper layers. However, using deep ANNs for training spiking models is still considered as a tedious task. Until recently, various ANN to SNN conversion methods in the literature have been proposed to train deep SNN models. Nevertheless, these methods require hundreds to thousands of time-steps for training and still cannot attain good SNN performance. This work proposes a customized model (VGG, ResNet) architecture to train deep convolutional spiking neural networks. In this current study, the training is carried out using deep convolutional spiking neural networks with surrogate gradient descent backpropagation in a customized layer architecture similar to deep artificial neural networks. Moreover, this work also proposes fewer time-steps for training SNNs with surrogate gradient descent. During the training with surrogate gradient descent backpropagation, overfitting problems have been encountered. To overcome these problems, this work refines the SNN based dropout technique with surrogate gradient descent. The proposed customized SNN models achieve good classification results on both private and public datasets. In this work, several experiments have been carried out on an embedded platform (NVIDIA JETSON TX2 board), where the deployment of customized SNN models has been extensively conducted. Performance validations have been carried out in terms of processing time and inference accuracy between PC and embedded platforms, showing that the proposed customized models and training techniques are feasible for achieving a better performance on various datasets such as CIFAR-10, MNIST, SVHN, and private KITTI and Korean License plate dataset.

Vision-Based Tactile Sensor Mechanism for the Estimation of Contact Position and Force Distribution Using Deep Learning.

This work describes the development of a vision-based tactile sensor system that utilizes the image-based information of the tactile sensor in conjunction with input loads at various motions to train the neural network for the estimation of tactile contact position, area, and force distribution. The current study also addresses pragmatic aspects, such as choice of the thickness and materials for the tactile fingertips and surface tendency, etc. The overall vision-based tactile sensor equipment interacts with an actuating motion controller, force gauge, and control PC (personal computer) with a LabVIEW software on it. The image acquisition was carried out using a compact stereo camera setup mounted inside the elastic body to observe and measure the amount of deformation by the motion and input load. The vision-based tactile sensor test bench was employed to collect the output contact position, angle, and force distribution caused by various randomly considered input loads for motion in X, Y, Z directions and RxRy rotational motion. The retrieved image information, contact position, area, and force distribution from different input loads with specified 3D position and angle are utilized for deep learning. A convolutional neural network VGG-16 classification modelhas been modified to a regression network model and transfer learning was applied to suit the regression task of estimating contact position and force distribution. Several experiments were carried out using thick and thin sized tactile sensors with various shapes, such as circle, square, hexagon, for better validation of the predicted contact position, contact area, and force distribution.

Surface Thermo-Dynamic Characterization of Poly (Vinylidene Chloride-Co-Acrylonitrile) (P(VDC-co-AN)) Using Inverse-Gas Chromatography and Investigation of Visual Traits Using Computer Vision Image Processing Algorithms.

The Inverse Gas Chromatography (IGC) technique has been employed for the surface thermo-dynamic characterization of the polymer Poly(vinylidene chloride-co-acrylonitrile) (P(VDC-co-AN)) in its pure form. IGC attributes, such as London dispersive surface energy, Gibbs free energy, and Guttman Lewis acid-base parameters were analyzed for the polymer (P(VDC-co-AN)). The London dispersive surface free energy ( γ S L ) was calculated using the Schultz and Dorris-Gray method. The maximum surface energy value of (P(VDC-co-AN )) is found to be 29.93 mJ·m - 2 and 24.15 mJ·m - 2 in both methods respectively. In our analysis, it is observed that the γ S L values decline linearly with an increase in temperature. The Guttman-Lewis acid-base parameter K a , K b values were estimated to be 0.13 and 0.49. Additionally, the surface character S value and the correlation coefficient were estimated to be 3.77 and 0.98 respectively. After the thermo-dynamic surface characterization, the (P(VDC-co-AN)) polymer overall surface character is found to be basic. The substantial results revealed that the (P(VDC-co-AN)) polymer surface contains more basic sites than acidic sites and, hence, can closely associate in acidic media. Additionally, visual traits of the polymer (P(VDC-co-AN)) were investigated by employing Computer Vision and Image Processing (CVIP) techniques on Scanning Electron Microscopy (SEM) images captured at resolutions ×50, ×200 and ×500. Several visual traits, such as intricate patterns, surface morphology, texture/roughness, particle area distribution ( D A ), directionality ( D P ), mean average particle area ( µ a v g ) and mean average particle standard deviation ( σ a v g ), were investigated on the polymer's purest form. This collective study facilitates the researches to explore the pure form of the polymer Poly(vinylidene chloride-co-acrylonitrile) (P(VDC-co-AN )) in both chemical and visual perspective.

Automatic Distortion Rectification of Wide-Angle Images Using Outlier Refinement for Streamlining Vision Tasks.

The study proposes an outlier refinement methodology for automatic distortion rectification of wide-angle and fish-eye lens camera models in the context of streamlining vision-based tasks. The line-members sets are estimated in a scene through accumulation of line candidates emerging from the same edge source. An iterative optimization with an outlier refinement scheme was applied to the loss value, to simultaneously remove the extremely curved outliers from the line-members set and update the robust line members as well as estimating the best-fit distortion parameters with lowest possible loss. The proposed algorithm was able to rectify the distortions of wide-angle and fish-eye cameras even in extreme conditions such as heavy illumination changes and severe lens distortions. Experiments were conducted using various evaluation metrics both at the pixel-level (image quality, edge stretching effects, pixel-point error) as well as higher-level use-cases (object detection, height estimation) with respect to real and synthetic data from publicly available, privately acquired sources. The performance evaluations of the proposed algorithm have been investigated using an ablation study on various datasets in correspondence to the significance analysis of the refinement scheme and loss function. Several quantitative and qualitative comparisons were carried out on the proposed approach against various self-calibration approaches.

Spine Computed Tomography to Magnetic Resonance Image Synthesis Using Generative Adversarial Networks : A Preliminary Study.

OBJECTIVE: To generate synthetic spine magnetic resonance (MR) images from spine computed tomography (CT) using generative adversarial networks (GANs), as well as to determine the similarities between synthesized and real MR images. METHODS: GANs were trained to transform spine CT image slices into spine magnetic resonance T2 weighted (MRT2) axial image slices by combining adversarial loss and voxel-wise loss. Experiments were performed using 280 pairs of lumbar spine CT scans and MRT2 images. The MRT2 images were then synthesized from 15 other spine CT scans. To evaluate whether the synthetic MR images were realistic, two radiologists, two spine surgeons, and two residents blindly classified the real and synthetic MRT2 images. Two experienced radiologists then evaluated the similarities between subdivisions of the real and synthetic MRT2 images. Quantitative analysis of the synthetic MRT2 images was performed using the mean absolute error (MAE) and peak signal-to-noise ratio (PSNR). RESULTS: The mean overall similarity of the synthetic MRT2 images evaluated by radiologists was 80.2%. In the blind classification of the real MRT2 images, the failure rate ranged from 0% to 40%. The MAE value of each image ranged from 13.75 to 34.24 pixels (mean, 21.19 pixels), and the PSNR of each image ranged from 61.96 to 68.16 dB (mean, 64.92 dB). CONCLUSION: This was the first study to apply GANs to synthesize spine MR images from CT images. Despite the small dataset of 280 pairs, the synthetic MR images were relatively well implemented. Synthesis of medical images using GANs is a new paradigm of artificial intelligence application in medical imaging. We expect that synthesis of MR images from spine CT images using GANs will improve the diagnostic usefulness of CT. To better inform the clinical applications of this technique, further studies are needed involving a large dataset, a variety of pathologies, and other MR sequence of the lumbar spine.

Feasible Self-Calibration of Larger Field-of-View (FOV) Camera Sensors for the Advanced Driver-Assistance System (ADAS).

This paper proposes a self-calibration method that can be applied for multiple larger field-of-view (FOV) camera models on an advanced driver-assistance system (ADAS). Firstly, the proposed method performs a series of pre-processing steps such as edge detection, length thresholding, and edge grouping for the segregation of robust line candidates from the pool of initial distortion line segments. A novel straightness cost constraint with a cross-entropy loss was imposed on the selected line candidates, thereby exploiting that novel loss to optimize the lens-distortion parameters using the Levenberg-Marquardt (LM) optimization approach. The best-fit distortion parameters are used for the undistortion of an image frame, thereby employing various high-end vision-based tasks on the distortion-rectified frame. In this study, an investigation was carried out on experimental approaches such as parameter sharing between multiple camera systems and model-specific empirical γ -residual rectification factor. The quantitative comparisons were carried out between the proposed method and traditional OpenCV method as well as contemporary state-of-the-art self-calibration techniques on KITTI dataset with synthetically generated distortion ranges. Famous image consistency metrics such as peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and position error in salient points estimation were employed for the performance evaluations. Finally, for a better performance validation of the proposed system on a real-time ADAS platform, a pragmatic approach of qualitative analysis has been conducted through streamlining high-end vision-based tasks such as object detection, localization, and mapping, and auto-parking on undistorted frames.

Deep CT to MR Synthesis Using Paired and Unpaired Data.

Magnetic resonance (MR) imaging plays a highly important role in radiotherapy treatment planning for the segmentation of tumor volumes and organs. However, the use of MR is limited, owing to its high cost and the increased use of metal implants for patients. This study is aimed towards patients who are contraindicated owing to claustrophobia and cardiac pacemakers, and many scenarios in which only computed tomography (CT) images are available, such as emergencies, situations lacking an MR scanner, and situations in which the cost of obtaining an MR scan is prohibitive. From medical practice, our approach can be adopted as a screening method by radiologists to observe abnormal anatomical lesions in certain diseases that are difficult to diagnose by CT. The proposed approach can estimate an MR image based on a CT image using paired and unpaired training data. In contrast to existing synthetic methods for medical imaging, which depend on sparse pairwise-aligned data or plentiful unpaired data, the proposed approach alleviates the rigid registration of paired training, and overcomes the context-misalignment problem of unpaired training. A generative adversarial network was trained to transform two-dimensional (2D) brain CT image slices into 2D brain MR image slices, combining the adversarial, dual cycle-consistent, and voxel-wise losses. Qualitative and quantitative comparisons against independent paired and unpaired training methods demonstrated the superiority of our approach.

fPADnet: Small and Efficient Convolutional Neural Network for Presentation Attack Detection.

The rapid growth of fingerprint authentication-based applications makes presentation attack detection, which is the detection of fake fingerprints, become a crucial problem. There have been numerous attempts to deal with this problem; however, the existing algorithms have a significant trade-off between accuracy and computational complexity. This paper proposes a presentation attack detection method using Convolutional Neural Networks (CNN), named fPADnet (fingerprint Presentation Attack Detection network), which consists of Fire and Gram-K modules. Fire modules of fPADnet are designed following the structure of the SqueezeNet Fire module. Gram-K modules, which are derived from the Gram matrix, are used to extract texture information since texture can provide useful features in distinguishing between real and fake fingerprints. Combining Fire and Gram-K modules results in a compact and efficient network for fake fingerprint detection. Experimental results on three public databases, including LivDet 2011, 2013 and 2015, show that fPADnet can achieve an average detection error rate of 2.61%, which is comparable to the state-of-the-art accuracy, while the network size and processing time are significantly reduced.

An in vitro comparative study of T2 and T2* mappings of human articular cartilage at 3-Tesla MRI using histology as the standard of reference.

OBJECTIVE: The aim of this study was to evaluate the correlations between T2 value, T2* value, and histological grades of degenerated human articular cartilage. MATERIALS AND METHODS: T2 mapping and T2* mapping of nine tibial osteochondral specimens were obtained using a 3-T MRI after total knee arthroplasty. A total of 94 ROIs were analyzed. Histological grades were assessed using the David-Vaudey scale. Spearman's rho correlation analysis and Pearson's correlation analysis were performed. RESULTS: The mean relaxation values in T2 map with different histological grades (0, 1, 2) of the cartilage were 51.9 ± 9.2 ms, 55.8 ± 12.8 ms, and 59.6 ± 10.2 ms, respectively. The mean relaxation values in T2* map with different histological grades (0, 1, 2) of the cartilage were 20.3 ± 10.3 ms, 21.1 ± 12.4 ms, and 15.4 ± 8.5 ms, respectively. Spearman's rho correlation analysis confirmed a positive correlation between T2 value and histological grade (ρ = 0.313, p < 0.05). Pearson's correlation analysis revealed a significant negative correlation between T2 and T2* (r = -0.322, p < 0.05). Although T2* values showed a decreasing trend with an increase in cartilage degeneration, this correlation was not statistically significant in this study (ρ = -0.192, p = 0.129). CONCLUSIONS: T2 mapping was correlated with histological degeneration, and it may be a good biomarker for osteoarthritis in human articular cartilage. However, the strength of the correlation was weak (ρ = 0.313). Although T2* values showed a decreasing trend with an increase in cartilage degeneration, the correlation was not statistically significant. Therefore, T2 mapping may be more appropriate for the initial diagnosis of articular cartilage degeneration in the knee joint. Further studies on T2* mapping are needed to confirm its reliability and mechanism in cartilage degeneration.