ABSTRACT
Kolam, the art of floor drawing is a long standing and time-honoured practice in India. The traditional art has its own impact in Indian culture because of its role in religious practice. The culture of kolam has decreased over time due to modernization. The photograph of these artistic images is the fundamental component that can provide the visual and sensory experiences to future generations. So, we need a sincere attempt to preserve and document the culture of floor art images. Image classification aids in easing the process of preservation and documentation of kolam images. As a primary step, we aim to classify the kolam images viz, footprint, swastik, geometric, plant and animal motifs in this article. Inspired by the performance of deep networks in most of the vision-based problems, we developed a new kolamNet to classify the different motifs of kolam. KolamNet is attained by incorporating attention mechanism into EfficientNet. Attention mechanism assures lightweight functionality and refines the deep feature along with end to end training. To validate KolamNet, Kolam dataset is created to accomplish the classification task. Extensive experimentation is conducted on Kolam dataset to prove the effectiveness of the KolamNet to classify the motifs. KolamNet shows better capability in terms of performance metrics – accuracy-97%, precision-0.96, recall-0.97 and F1score-0.97 when compared to related state of the art deep network architectures.
- Haikel Alhichri, Asma S Alswayed, Yakoub Bazi, Nassim Ammour, and Naif A Alajlan. 2021. Classification of remote sensing images using EfficientNet-B3 CNN model with attention. IEEE Access 9(2021), 14078–14094.Google ScholarCross Ref
- M Ascher. 2002. The Kolam Tradition: A tradition of figure-drawing in southern India expresses mathematical ideas and has attracted the attention of computer science. American Scientist 1(2002), 56–63.Google ScholarCross Ref
- U M Ascher and L R Petzold. 1991. Projected implicit Runge–Kutta methods for differential-algebraic equations. SIAM J. Numer. Anal. 4, 28 (1991), 1097–1120.Google ScholarDigital Library
- Won Hee Chung, Yeong Hyeon Gu, and Seong Joon Yoo. 2022. District heater load forecasting based on machine learning and parallel CNN-LSTM attention. Energy (Oxf.) 246, 123350 (May 2022), 123350.Google ScholarCross Ref
- A S Ebenezer, S D Kanmani, M Sivakumar, and S J Priya. 2022. Effect of image transformation on EfficientNet model for COVID-19 CT image classification. Materials Today: Proceedings 51 (2022), 2512–2519.Google ScholarCross Ref
- Anna Laine. 2013. Kolam patterns as materialisation and embodiment of rhythms. Anthrovision1.2 (April 2013).Google Scholar
- John Layard. 1937. Labyrinth ritual in south India: Threshold and tattoo designs. Folklore 48, 2 (June 1937), 115–182.Google ScholarCross Ref
- Anika Sarin. 2022. The Kolam drawing: A point lattice system. Des. issues 38, 3 (July 2022), 34–54.Google Scholar
- Gift Siromoney and Rani Siromoney. 1987. Rosenfeld’s cycle grammars and kolam. In Lecture Notes in Computer Science. Springer Berlin Heidelberg, Berlin, Heidelberg, 564–579.Google Scholar
- Mingxing Tan and Quoc V Le. 2019. EfficientNet: Rethinking model scaling for convolutional Neural Networks. (May 2019). arxiv:1905.11946 [cs.LG]Google Scholar
- Vibhor Kumar Vishnoi, Krishan Kumar, and Brajesh Kumar. 2021. Plant disease detection using computational intelligence and image processing. J. Plant Dis. Prot. (2006) 128, 1 (Feb. 2021), 19–53.Google Scholar
- Ab Wahab, M N Nazir, A Ren, A T Z Noor, M H M Akbar, and M F Mohamed. 2021. Efficientnet-lite and hybrid CNN-KNN implementation for facial expression recognition on raspberry pi. IEEE Access 9(2021), 134065–134080.Google ScholarCross Ref
- Sanghyun Woo, Jongchan Park, Joon-Young Lee, and In So Kweon. 2018. CBAM: Convolutional Block Attention Module. (July 2018). arxiv:1807.06521 [cs.CV]Google Scholar
- K Yanagisawa and S Nagata. 2007. Fundamental study on design system of kolam pattern. Forma 22(2007), 31–46.Google Scholar
- Peng Zhao, Yang Liu, Yijuan Lu, and Benpeng Xu. 2019. A sketch recognition method based on transfer deep learning with the fusion of multi-granular sketches. Multimed. Tools Appl. 78, 24 (Dec. 2019), 35179–35193.Google ScholarCross Ref
Index Terms
- KolamNet: An attention based model for kolam classification✱
Recommendations
Exploring Kolam As An Ecofeminist Computational Art Practice
C&C '22: Proceedings of the 14th Conference on Creativity and CognitionIn this pictorial, we present Kolam, a visual artform originating in Tamilnadu, South India, as an ecofeminist computational art practice. We provide a visual documentation of Kolam's Intangible Cultural Heritage (ICH) through eight characteristics ...
Genome wide classification and characterisation of CpG sites in cancer and normal cells
This study identifies common methylation patterns across different cancer types in an effort to identify common molecular events in diverse types of cancer cells and provides evidence for the sequence surrounding a CpG to influence its susceptibility to ...
Generating various kolam patterns using new kolam picture grammar
Kolam is an artistic creation. It is a ubiquitous art form predominant in South India, while also seen in a few places in northern India and South East Asia. Kolam holds a rich tradition of cultural and medicinal significance. Kolams are generated using ...
Comments