The proper handling of waste is one of the biggest challenges of modern society. Municipal Solid Waste (MSW) requires categorization into a number of types, including bio, plastic, glass, metal, paper, etc. The most efficient techniques proposed by researchers so far include neural networks. In this paper, a detailed summarization was made of the existing deep learning techniques that have been proposed to classify waste. This paper proposes an architecture for the classification of litter into the categories specified in the benchmark approaches. The architecture used for classification was EfficientNet-B0. These are compound-scaling based models proposed by Google that are pretrained on ImageNet and have an accuracy of 74% to 84% in top-1 over ImageNet. This research proposes EfficientNet-B0 model tuning for images specific to particular demographic regions for efficient classification. This type of model tuning over transfer learning provides a customized model for classification, highly optimized for a particular region. It was shown that such a model had comparable accuracy to that of EfficientNet-B3, however, with a significantly smaller number of parameters required by the B3 model. Thus, the proposed technique achieved efficiency on the order of 4X in terms of FLOPS. Moreover, it resulted in improvised classifications as a result of fine-tuning over region-wise specific litter images.

In cloud computing, a shared, configurable pool of computing resources is made accessible online and allocated to users according to their needs. It is essential for a cloud provider to schedule jobs in the cloud to keep up service quality and boost system efficiency. In this paper, we present a scheduling technique based on modified particle swarm optimization to combat the issues of excessively long scheduling time and high computation costs associated with scheduling jobs in a cloud environment. The modified PSO is used to allocate the jobs to virtual machines in order to minimize the objective function consisting of cost and makespan. The algorithm relies on biological changes that occur in organisms to regulate premature convergence and improve local search capability. The technique is analyzed and simulated using CloudSim, and the simulation results demonstrate that the proposed approach decreases makespan and cost effectively as compared to standard PSO.

Watermarking of digital images is required in diversified applications ranging from medical imaging to commercial images used over the web. Usually, the copyright information is embossed over the image in the form of a logo at the corner or diagonal text in the background. However, this form of visible watermarking is not suitable for a large class of applications. In all such cases, a hidden watermark is embedded inside the original image as proof of ownership. A large number of techniques and algorithms are proposed by researchers for invisible watermarking. In this paper, we focus on issues that are critical for security aspects in the most common domains like digital photography copyrighting, online image stores, etc. The requirements of this class of application include robustness (resistance to attack), blindness (direct extraction without original image), high embedding capacity, high Peak Signal to Noise Ratio (PSNR), and high Structural Similarity Matrix (SSIM). Most of these requirements are conflicting, which means that an attempt to maximize one requirement harms the other. In this paper, a blind type of image watermarking scheme is proposed using Lifting Wavelet Transform (LWT) as the baseline. Using this technique, custom binary watermarks in the form of a binary string can be embedded. Hu’s Invariant moments’ coefficients are used as a key to extract the watermark. A Stochastic variant of the Firefly algorithm (FA) is used for the optimization of the technique. Under a prespecified size of embedding data, high PSNR and SSIM are obtained using the Stochastic Gradient variant of the Firefly technique. The simulation is done using Matrix Laboratory (MATLAB) tool and it is shown that the proposed technique outperforms the benchmark techniques of watermarking considering PSNR and SSIM as quality metrics.

The standard of healthy intake of food is the necessity for keeping a balanced diet to prevent the obesity problem and many other health problems in humans. Obesity is increasing at an alarming speed and keeping people's health at risk. Mankind needs to have careful control on their daily intake of calories by choosing healthier foods, which will be the most fundamental method in preventing obesity and ill health. Even though the packaging of food comes with calorie and nutrition labels, it might not be very favorable for the reference of people. Thus, the scientists to help people started using AI-based techniques and methodologies to know the ways of determining their daily calorie intake of their food. This chapter proposes a review of various AI-based food calorie estimation methodologies in diet assessment which are suggested to help the normal people and patients so that normal people and doctors could succeed to fight against diet-based health conditions.