Search

Abstract Kolkata, renowned as the City of Joy, boasts a rich tapestry of cultural heritage spanning centuries. Despite the significance of its architectural marvels, accessing comprehensive visual documentation of Kolkata's heritage sites remains a challenge. In online searches, limited imagery often fails to provide a detailed understanding of these historical landmarks. To address this gap, this paper introduces MonuNet, a high-performance deep-learning network specifically designed for the classification of heritage images from Kolkata. The development of MonuNet addresses the critical need for efficient and accurate identification of Kolkata's architectural marvels, which are significant tangible cultural heritages. The dataset used to train MonuNet is organized by heritage sites, each category within the dataset represents distinct sites. It includes images from 13 prominent heritage sites in Kolkata. For each of these sites, there are 50 images, making it a structured collection where each category (heritage site) is equally represented. The proposed network utilizes a unique architecture incorporating a Dense channel attention module and a Parallel-spatial channel attention module to capture intricate architectural details and spatial relationships within the images. Experimental evaluations demonstrate the superior performance of MonuNet in classifying Kolkata heritage images with an accuracy of 89%, Precision of 87.77%, and Recall of 86.61%. The successful deployment of MonuNet holds significant implications for cultural preservation, tourism enhancement, and urban planning in Kolkata, aligning with the United Nations Sustainable Development Goals (SDGs) for sustainable city development. By providing a robust tool for the automatic identification and classification of heritage images, MonuNet promises to enrich online repositories with detailed visual documentation, thereby enhancing accessibility to Kolkata's cultural heritage for researchers, tourists, and urban planners alike. Graphical Abstract

Abstract In India, kolam, commonly referred to as rangoli, is a traditional style of art. It involves using rice flour, chalk, or coloured powders to create elaborate patterns and motifs on the ground. Kolam is a common daily ritual in many regions of India, especially in South India, where it is seen as a significant cultural tradition and a means to greet visitors. Unfortunately, as a result of people’s hectic lives nowadays, the habit of drawing kolam on a regular basis is dwindling. The art of making kolam patterns is in danger of disappearing as so many individuals no longer have the time or space to do it on a regular basis. Therefore, it is imperative that ancient art be conserved and digitally documented in order to enlighten our next generation about kolam and its classifications. Deep learning has become a powerful technique because of its ability to learn from raw image data without the aid of a feature engineering process. In this article, we attempted to understand the types of Kolam images using the proposed deep architecture called KolamNetV2. KolamNetV2 comprises EfficientNet and attention layers, ensuring high accuracy with minimal training data and parameters. We evaluated KolamNetV2 to reveal its ability to learn the various types in our challenging Kolam dataset. The experimental findings show that the proposed network achieves fine enhancement in performance metrics viz, precision-0.7954, recall-0.7846, F1score-0.7854 and accuracy-81%. We compared our results with state-of-the-art deep learning methodologies, proving the astounding capability. Graphical Abstract

This article summarizes the Detection of Pond Fish Challenge (DePondfi’23 Challenge), held during the National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics (NCVPRIPG 2023). The main goal of the challenge was to find the most effective methods for detecting pond fish in underwater images, overcoming obstacles such as poor visibility, variations in turbidity, and environmental shifts. Sixty participants registered, with 15 teams submitting results for phase 1. The challenge concluded with four teams earning top honors based on mAP (mean Average Precision) score and time complexity. The mAP scores achieved by toppers are as follows: DETECTRON - 38.93%, DMACS SAI - 36.65%, PondVision - 31.63%, and Sahajeevis - 29.06%. This article describes the toppers method and discusses the detection results. Our challenge event is in line with Sustainable Development Goal 14, which focuses on the conservation and sustainable utilization of ponds and marine resources for sustainable development.

Sports employ computer vision for a variety of applications, including interactive viewing, object interpretation, sports action analysis, and intelligent rule-based systems for better referee choices. The computer vision problem known as “sports action recognition” looks for interactions between players and related sporting items. Most of the existing solutions in sports action recognition are complex and incomplete as taking both playfield and non-play field objects with various actions for the same human posture into account. Hence, the proposed model focuses only on objects within the playfield and the right functional action that is performed by the human. We identify the play field using a quadrant-based density (QD) approach which locates the crowded region and then morphological operations are applied to spot only play region. The right action performed is identified using the logic-based query to the Sports Action Generation model (SAG) which is generated using the newly constructed sports ontology. The proposed semantic verification model using ontology typically offers more transparency compared to deep learning networks. It relies on explicit rules and logic that can be easily interpreted, enhancing interoperability and accuracy across systems. We use the sports video wild (SVW) and UCF-101 datasets to assess the performance of our proposed model. The proposed sports action recognition system achieves action identification in videos with a good accuracy of 96%.

Busbars are essential components in electrical power distribution but are susceptible to overheating under high loads, potentially leading to system failures and equipment damage. Traditional overheating detection methods are often labour-intensive and require system shutdowns. To address these issues, we developed a smart, non-destructive system utilising thermal imaging for continuous monitoring of busbars. Thermal imaging offers non-invasive detection of temperature anomalies without interrupting operations, enabling early identification of overheating and supporting preventative maintenance. Our system combines the Hjorth thermal pattern extraction method with a novel one-dimensional convolutional neural network, BusThermNet. A new dataset of thermal images capturing busbar conditions at low and high loads has been created for this study. The Hjorth method analyzes busbar dynamics by evaluating activity, mobility, and complexity, while BusThermNet classifies the data into low- and high-load categories. Upon validation with the new dataset, our approach achieved an accuracy of 82.47%, with a precision of 0.89, recall of 0.74, and an F1 score of 0.81, outperforming existing techniques. These results highlight the potential of our system as a valuable tool for electrical regulatory authorities. Its capabilities for early issue detection, real-time monitoring, and reduced system disruptions make it an effective proactive solution for busbar management, ideal for adoption by regulatory bodies and industry stakeholders. [ABSTRACT FROM AUTHOR]