Plastic bottle recycling has a crucial role in environmental degradation and protection. Position and background should be the same to classify plastic bottles on a conveyor belt. The manual detection of plastic bottles is time consuming and leads to human error. Hence, the automatic classification of plastic bottles using deep learning techniques can assist with the more accurate results and reduce cost. To achieve a considerably good result using the DL model, we need a large volume of data to train. We propose a GAN-based model to generate synthetic images similar to the original. To improve the image synthesis quality with less training time and decrease the chances of mode collapse, we propose a modified lightweight-GAN model, which consists of a generator and a discriminator with an auto-encoding feature to capture essential parts of the input image and to encourage the generator to produce a wide range of real data. Then a newly designed weighted average ensemble model based on two pre-trained models, inceptionV3 and xception, to classify transparent plastic bottles obtains an improved classification accuracy of 99.06%.

Nowadays user devices such as phones, tablets etc. allows processing the images with help of high-end applications and softwares developed. Most of the times, the images are downscaled to make them compatible with these end devices. This leads to the loss of image quality. This loss of information on downscaling an image results in distortion of edges and while zoomed in results into a blurred image. As the edge detection is a basic step for many image processing applications such as object detection, object segmentation, object recognition, etc. It is necessary to know the impact of edge detection on downscaled image. In this paper, we are using Canny Edge detection method to detect the edges. The original images are downscaled using different interpolation methods. Canny Edge detection is applied on original images and downscaled images to compare the distortion in the edges. We used Structural Similarity Index Method for comparison. We are also comparing execution time taken by Canny Edge Detection on different interpolation methods to check for optimal interpolation method. We observed that the distortion in edges and time efficiency differ for different interpolation methods which are detailed below in the result section. As blurring is also a disadvantage of downscaling, we are applying Gaussian Blur on the images to compare the blurring due to Gaussian blur technique and blurring due to downscaling.

Image segmentation plays a central role in a broad range of applications, such as medical image analysis, autonomous vehicles, video surveillance and augmented reality. Portrait segmentation, which is a subset of semantic image segmentation, is widely used as a preprocessing step in multiple applications such as security systems, entertainment applications, video conferences, etc. A substantial amount of deep learning-based portrait segmentation approaches have been developed, since the performance and accuracy of semantic image segmentation have improved significantly due to the recent introduction of deep learning technology. However, these approaches are limited to a single portrait segmentation model. In this paper, we propose a novel approach using an ensemble method by combining multiple heterogeneous deep-learning based portrait segmentation models to improve the segmentation performance. The Two-Models ensemble and Three-Models ensemble, using a simple soft voting method and weighted soft voting method, were experimented. Intersection over Union (IoU) metric, IoU standard deviation and false prediction rate were used to evaluate the performance. Cost efficiency was calculated to analyze the efficiency of segmentation. The experiment results show that the proposed ensemble approach can perform with higher accuracy and lower errors than single deep-learning-based portrait segmentation models. The results also show that the ensemble of deep-learning models typically increases the use of memory and computing power, although it also shows that the ensemble of deep-learning models can perform more efficiently than a single model with higher accuracy using less memory and less computing power.

With the rapid increase in world’s population and changing climate patterns, accurate crop yield forecasting is essential to ensure food security and sustainable agriculture. This study presents a yield prediction framework consisting of Stacking Ensemble Model (SEM) and its Optimized variant (OSEM), which integrates real-world agricultural data with synthetic data generated using the Prophet time-series model. The ensemble comprises Random Forest, XGBoost, Decision Tree, and K-Nearest Neighbors as base learners, with an Extra Trees Regressor as the meta-learner. OSEM incorporates hyperparameter tuning and dimensionality reduction using Singular Value Decomposition to enhance predictive performance and reduce feature noise. Extensive experiments on globally distributed and region-specific crop datasets demonstrate that OSEM consistently outperforms both SEM and state-of-the-art baselines, achieving $R^{2} = 0.996$ and MAE = 0.185 t/ha. Stability and convergence analysis, supported by Wilcoxon and Friedman tests, further confirm the robustness and reliability of the proposed model for scalable crop yield forecasting.

The dumbbell (DB) ZnO bipod crystal is directly grown from acetate anion intercalated Zn based hydroxyl double salts (Zn-HDS). Our results suggest that the (0001)-to-(0001) plane orientation of the DB ZnO crystal comes from the acetate anions intercalated into the Zn-HDS. The DB ZnO crystal is dynamically changed to tube, nanoplate, and the new Zn-HDS with interlayer spacing of 20.0Å by etching the (000-1) surface.