Development of Scalable Coding of Encrypted Images Using Modified Absolute Moment Block Truncation Code

Many researchers worked on scalable coding for unencrypted images, and there is more space for research in scalable coding for encrypted images. This paper proposes a novel method of scalable coding for encrypted images, especially for lossy compression images using the Modified Absolute Moment Block Truncation Code (MAMBTC) technique. The given input image is compressed using MAMBTC and then encrypted using a Pseudo-Random Number (PRNG) at the encryption phase. The PRNG is shared between the encoder and the decoder. At the decryption phase, the compressed pixel value is obtained by decryption using the PRNG and then reconstructed using MAMBTC, scaled by scaling factor 2 and Bilinear Interpolation Technique to obtain the original image. MAMBTC gives better image quality than Block Truncation Code (BTC), a higher PSNR of 36.32 dB, and a Compression ratio of 1.09, which makes the proposed system ready for the signal processing community/applications.

Multimedia Concepts on Object Detection and Recognition with F1 Car Simulation Using Convolutional Layers

This paper presents a framework for detecting objects in images based on global features and contours. The first step is a shape matching algorithm that uses the background subtraction process. Object detection is accomplished by an examination of the oversegmentation of the image, where the space of the potential boundary of the object is examined to identify boundaries that have a direct resemblance to the prototype of the object type to be detected. Our analysis method removes edges using bilinear interpolation and reestablishes color sensors as lines and retracts background lines from the previous frame. Object contours are generated with clustered lines. The objects detected will then be recognized using the extraction technique. Here, we analyze the color and shape characteristics with which each object is capable of managing occlusion and interference. As an extension of object detection and recognition, F1 car simulation is experimented with simulation using various layers, such as layer drops, convolutionary layers, and boundary elimination, avoiding obstacles in different pathways.

Comprehensive Review & Enhancements on Probe Substitution Techniques for Faulty Steady State Inlet Pressure Distortion Data

When a gas-turbine engine is in operation, inlet-generated total-pressure distortion can have a detrimental effect on engine’s stability and performance. During the product development life cycle, on-ground wind tunnel tests and in-flight tests are performed to estimate the inlet distortion characteristics. Extensive measures are taken in the preparation and execution of inlet distortion tests. The data pertaining to spatial inlet distortion is recorded using an array of high-response total-pressure probes. The pressure probes are usually arranged in rake and ring arrays as per AIR1419. The data from these probes is used by propulsion system designers to address the effects of inlet distortion on stability and performance, particularly the engine’s sensitivity to inlet distortion. In some instances, the probes can produce inaccurate measurements or no measurements at all, due to a variety of reasons. This may result in a time consuming and costly process of repeating the test. To avoid this, the inaccurate or invalid measurements can be substituted using a variety of statistical techniques during test data post-processing. This paper discusses the results of different interpolation techniques to substitute invalid steady-state total-pressure measurements, evaluated in the context of classical distortion profile data available in AIR1419. The techniques include 1D linear interpolation using only probes data from adjacent rings, 1D linear interpolation using only probes data from adjacent rakes, and bilinear interpolation using probes data from adjacent rings and rakes. Furthermore, the paper evaluates a bilinear interpolation technique with optimal weights obtained from linear regression, that enhances the estimation of invalid pressure values.

Detection and Segmentation of Mature Green Tomatoes Based on Mask R-CNN with Automatic Image Acquisition Approach

Since the mature green tomatoes have color similar to branches and leaves, some are shaded by branches and leaves, and overlapped by other tomatoes, the accurate detection and location of these tomatoes is rather difficult. This paper proposes to use the Mask R-CNN algorithm for the detection and segmentation of mature green tomatoes. A mobile robot is designed to collect images round-the-clock and with different conditions in the whole greenhouse, thus, to make sure the captured dataset are not only objects with the interest of users. After the training process, RestNet50-FPN is selected as the backbone network. Then, the feature map is trained through the region proposal network to generate the region of interest (ROI), and the ROIAlign bilinear interpolation is used to calculate the target region, such that the corresponding region in the feature map is pooled to a fixed size based on the position coordinates of the preselection box. Finally, the detection and segmentation of mature green tomatoes is realized by the parallel actions of ROI target categories, bounding box regression and mask. When the Intersection over Union is equal to 0.5, the performance of the trained model is the best. The experimental results show that the F1-Score of bounding box and mask region all achieve 92.0%. The image acquisition processes are fully unobservable, without any user preselection, which are a highly heterogenic mix, the selected Mask R-CNN algorithm could also accurately detect mature green tomatoes. The performance of this proposed model in a real greenhouse harvesting environment is also evaluated, thus facilitating the direct application in a tomato harvesting robot.

Face Transferring on Webcam images using StyleGAN

In image processing technology, face transfer is broadly used for privacy protection, picture enhancement, and entertainment applications. Face transfer is the domain that maps one image into another image and extracts several features of the face from one person to morph that face to another person. This face transfer will carry the facial expressions also. This is also called face morph, face swap, etc. Here we propose StyleGAN technology using face transfer with the image to get high quality. In this StyleGAN contribute the bilinear interpolation and affine transformation. Bilinear interpolation is to remove the noise and increase the quality of images. Affine transformation is to supply the images with 2d warping to improve the image quantity. To upgrade the quality of the images with face transfer is adopted to increase the accuracy of the image quality after image transfer.

Neural Network-Oriented Big Data Model for Yoga Movement Recognition

The use of computer vision for target detection and recognition has been an interesting and challenging area of research for the past three decades. Professional athletes and sports enthusiasts in general can be trained with appropriate systems for corrective training and assistive training. Such a need has motivated researchers to combine artificial intelligence with the field of sports to conduct research. In this paper, we propose a Mask Region-Convolutional Neural Network (MR-CNN)- based method for yoga movement recognition based on the image task of yoga movement recognition. The improved MR-CNN model is based on the framework and structure of the region-convolutional network, which proposes a certain number of candidate regions for the image by feature extraction and classifies them, then outputs these regions as detected bounding boxes, and does mask prediction for the candidate regions using segmentation branches. The improved MR-CNN model uses an improved deep residual network as the backbone network for feature extraction, bilinear interpolation of the extracted candidate regions using Region of Interest (RoI) Align, followed by target classification and detection, and segmentation of the image using the segmentation branch. The model improves the convolution part in the segmentation branch by replacing the original standard convolution with a depth-separable convolution to improve the network efficiency. Experimentally constructed polygon-labeled datasets are simulated using the algorithm. The deepening of the network and the use of depth-separable network improve the accuracy of detection while maintaining the reliability of the network and validate the effectiveness of the improved MR-CNN.

Asymmetrical pythagorean-hodograph spline-based C4 continuous local corner smoothing method with jerk continuous feedrate scheduling along linear toolpath

In computer numerical control systems, linear segments, which are generated by computer-aided manufacturing software, are the most widely used toolpath format. Since the linear toolpath is discontinuous at the junction of two adjacent segments, the fluctuations on velocity, acceleration and jerk are inevitable. Local corner smoothing is widely used to address this problem. However, most existing methods use symmetrical splines to smooth the corners. When any one of the linear segments at the corner is short, to avoid overlap, the inserted spline will be micro, thereby increasing the curvature extreme of the spline and reducing the feedrate along it. In this article, the corners are smoothed by a 𝐶4 continuous asymmetric Pythagorean-hodograph (PH) spline. The curvature extreme of the proposed spline is investigated first, and 𝐾=2.5 is determined as the threshold to constarin the asymmetry of the spline. Then a two-step strategy is used to generate a blended toolpath composed of asymmetric PH splines and linear segments. In the first step, the PH splines at the corners are generated under the premise that the transition lengths do not exceed half of the length of the linear segments. In the second step, the splines at the corners are re-planned to reduce the curvature extremes, if the transition error does not reach the given threshold and there are extra linear trajectories on both sides of the spline trajectory. Finally, the bilinear interpolation method is applied to determine the critical points of the smoothed toolpath, and a jerk-continuous feedrate scheduling scheme is presented to interpolate the smoothed toolpath. Simulations show that, under the condition of not affecting the machining quality, the proposed method can improve the machining efficiency by 7.84% to 23.98% compared to 𝐺3 and 𝐺4 methods.

Gated Skip-Connection Network with Adaptive Upsampling for Retinal Vessel Segmentation

Segmentation of retinal vessels is a critical step for the diagnosis of some fundus diseases. Methods: To further enhance the performance of vessel segmentation, we propose a method based on a gated skip-connection network with adaptive upsampling (GSAU-Net). In GSAU-Net, a novel skip-connection with gating is first utilized in the extension path, which facilitates the flow of information from the encoder to the decoder. Specifically, we used the gated skip-connection between the encoder and decoder to gate the lower-level information from the encoder. In the decoding phase, we used an adaptive upsampling to replace the bilinear interpolation, which recovers feature maps from the decoder to obtain the pixelwise prediction. Finally, we validated our method on the DRIVE, CHASE, and STARE datasets. Results: The experimental results showed that our proposed method outperformed some existing methods, such as DeepVessel, AG-Net, and IterNet, in terms of accuracy, F-measure, and AUCROC. The proposed method achieved a vessel segmentation F-measure of 83.13%, 81.40%, and 84.84% on the DRIVE, CHASE, and STARE datasets, respectively.