Advancements in multimodal biometrics, which integrate multiple biometric traits, promise to enhance the accuracy and robustness of identification systems. This study focuses on improving multimodal biometric identification by using fingerprint and finger vein images as the primary traits. We utilized the “NUPT-FPV” dataset, which contains a substantial number of finger vein and fingerprint images, which significantly aided our research. Convolutional neural networks (CNNs), renowned for their efficacy in computer vision tasks, are used in our model to extract distinct discriminative features. Specifically, we incorporate three popular CNN architectures: ResNet, VGGNet, and DenseNet. We explore three fusion strategies used in security applications: early fusion, late fusion, and score-level fusion. Early fusion integrates raw images at the input layer of a single CNN, combining information at the initial stages. Late fusion, in contrast, merges features after individual learning from each CNN model. Score-level fusion employs weighted aggregation to combine scores from each modality, leveraging the complementary information they provide. We also use contrast limited adaptive histogram equalization (CLAHE) to enhance fingerprint contrast and vein pattern features, improving feature visibility and extraction. Our evaluation metrics include accuracy, equal error rate (EER), and ROC curves. The fusion of CNN architectures and enhancement methods shows promising performance in identifying multimodal biometrics, aiming to increase identification accuracy. The proposed model offers a reliable authentication system using multiple biometrics to verify identity.

Waste management is one of the many major challenges faced by all urban cities around the world. With the increase in population, the current mechanisms for waste collection and disposal are under strain. The waste management problem is a global challenge that requires a collaborative effort from different stakeholders. Moreover, there is a need to develop technology-based solutions besides engaging the communities and establishing novel policies. While there are several challenges in waste management, the collection of waste using the current infrastructure is among the top challenges. Waste management suffers from issues such as a limited number of collection trucks, different types of household and industrial waste, and a low number of dumping points. The focus of this paper is on utilizing the available waste collection transportation capacity to efficiently dispose of the waste in a time-efficient manner while maximizing toxic waste disposal. A novel knapsack-based technique is proposed that fills the collection trucks with waste bins from different geographic locations by taking into account the amount of waste and toxicity in the bins using IoT sensors. Using the Knapsack technique, the collection trucks are loaded with waste bins up to their carrying capacity while maximizing their toxicity. The proposed model was implemented in MATLAB, and detailed simulation results show that the proposed technique outperforms other waste collection approaches. In particular, the amount of high-priority toxic waste collection was improved up to 47% using the proposed technique. Furthermore, the number of waste collection visits is reduced in the proposed scheme as compared to the conventional method, resulting in the recovery of the equipment cost in less than a year.

The efficient collection of municipal solid waste poses a significant challenge for the prospective development of smart cities. Using Internet of Things (IoT) technology enables the detection of various kinds of waste-related information, facilitating the implementation of a comprehensive plan for efficient waste collection. In this paper, an innovative waste collection mechanism leveraging the IoT sensors and widely recognized Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), a robust multi-criteria decision analysis method is introduced. Recognized for its ability to yield optimal solutions amidst multiple influencing parameters, TOPSIS proves particularly advantageous when confronted with diverse factors affecting decision making. A TOPSIS-based algorithm for optimal route selection is developed to collect waste using metrics such as toxicity, volume, and time duration of the waste while minimizing the route distance. To demonstrate the efficiency of our proposed algorithm, a comparative analysis with existing algorithms that are dependent on and optimized for single parameters is conducted. Through rigorous evaluation, it is shown that our proposed algorithm reduces the total distance traveled to collect the waste from all the bins by $14\%$ by effectively considering multiple criteria and optimizing waste collection routes.

Consumer product recognition involves utilizing computer vision (CV) and machine learning (ML) systems for identifying and recognizing numerous consumer goods in video or image frames. This technology determines applications in retail, e-commerce, security systems, and inventory management. The purpose is to automate the detection of certain products or objects in visual data, streamlining several procedures and improving efficacy. Deep Learning (DL), implementing structures such as convolutional neural networks (CNNs), automates the detection and classification of consumer products in images or video frames. Trained on diverse datasets, these techniques learn intricate features, benefiting retail and e-commerce by enhancing inventory management and improving the shopping experience through automated product detection. This study develops a honey badger algorithm with ensemble learning for consumer product identification and classification (HBAEL-CPIC) technique. The goal of the HBAEL-CPIC technique is to detect and classify various kinds of consumer products accurately. To accomplish this, the HBAEL-CPIC technique employs a non-local means (NLM) filter to eradicate the occurrence of the noise. Next, the HBAEL-CPIC technique follows the capsule network (CapsNet) model for the feature extractor process. Meanwhile, the hyperparameter tuning of the CapsNet approach has been carried out using the HBA. At last, the ensemble of three models namely extreme learning machine (ELM), descriptive back propagated: radial basis function (DBRF), and extreme gradient boosting (XGBoost) are used for classification. To highlight the improved detection solution of the HBAEL-CPIC technique, a series of experiments were performed. The comprehensive comparison study of the HBAEL-CPIC technique exhibited a superior accuracy value of 96.56% over other DL approaches.