An Optimized Model for the Segmentation of the Ancient Temple Vimanas using FCN Network
Keywords:
Archaeology, segmentation, vimana, Fully Convolutional Network (FCN), hyper parameters, recall, precision, Dice correlation coefficientAbstract
An extensive collection of artifacts, antiquities that are historically and archaeologically significant monuments is housed in the Indian state of Karnataka. Tradition and culture are intricately linked. Karnataka boasts a multitude of Neolithic and Megalithic structures, which have withstood the test of time for millennia. These architectural marvels are remnants of esteemed ruling dynasties. They possess unique wonders characterized by their distinctive style, inherent sculptural and architectural qualities, technical prowess, vastness, and grandeur. Nevertheless, the current generation is ill-prepared to extract archaeological knowledge pertaining to empires or reigning dynasties of these ancient Karnataka temples under the instruction of archaeologists. Therefore, it is necessary to adopt a novel method to effectively deliver this vital information to the contemporary age through a suitable platform. Archaeologists have numerous intricate challenges due to the absence of reliable digital techniques for automatically segmenting Vimana. Automated segmentation of Vimana poses challenges due to the variability in image acquisition, intricate architectural designs, noise, time difficulties, and photographic artifacts. As per our knowledge techniques for segmentation have not been proposed in the literature for vimana segmentation. Our work introduces a optimized fully convolutional network (FCN) model designed specifically for the automated segmentation of Vimana. The suggested approach mitigates the variability of image noise and trains Fully Convolutional Network (FCN) models using images from our custom dataset. Additionally, it has been demonstrated that employing appropriate data augmentation and model hyper-parameterization effectively mitigates over-fitting in the context of vimana area segmentation. The proposed methodology is evaluated using the test dataset, attaining a rate of recall of 0.9302 and a precision rate of 0.8977. The recommended method outperforms four other methods with lower depths in the segmentation problem, earning a Dice correlation coefficient of 0.8894 & with very min loss of around 0.1106. Finally a comparison of same methods with & without edge-smoothing is carried out. An improvement of 12%, 28% is achieved in DICE & PRECISION by an optimized FCN(U-Net) for the segmentation of vimana.
Downloads
References
Senthilkumaran, N., & Rajesh, R. (2009). A study on rough set theory for medical image segmentation. International journal of recent trends in Engineering, 2(2), 236.
Panwar, P., Gopal, G., & Kumar, R. (2016). Image Segmentation using K-means clustering and Thresholding. Image, 3(05), 1787-1793.
Panda, S. (2015). Color image segmentation using K-means clustering and thresholding technique. Interntional journal of ESC, 1132-1136.
Taneja, A., Ranjan, P., & Ujjlayan, A. (2015, September). A performance study of image segmentation techniques. In 2015 4th International Conference on Reliability, Infocom Technologies and Optimization (ICRITO)(Trends and Future Directions) (pp. 1-6). IEEE.
Cerimele, M. M., & Cossu, R. (2007). Decay regions segmentation from color images of ancient monuments using fast marching method. Journal of Cultural Heritage, 8(2), 170-175.
Masiero, A., Guarnieri, A., Pirotti, F., & Vettore, A. (2015). Semi-automated detection of surface degradation on bridges based on a level set method. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 40, 15-21.
Kamnitsas, K., Chen, L., Ledig, C., Rueckert, D., & Glocker, B. (2015). Multi-scale 3D convolutional neural networks for lesion segmentation in brain MRI. Ischemic stroke lesion segmentation, 13, 46.
Yu, L., Chen, H., Dou, Q., Qin, J., & Heng, P. A. (2016). Automated melanoma recognition in dermoscopy images via very deep residual networks. IEEE transactions on medical imaging, 36(4), 994-1004.
Ronneberger, O., Fischer, P., & Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18 (pp. 234-241). Springer International Publishing.
Long, J., Shelhamer, E., & Darrell, T. (2015). Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 3431-3440).
Milletari, F., Navab, N., & Ahmadi, S. A. (2016, October). V-net: Fully convolutional neural networks for volumetric medical image segmentation. In 2016 fourth international conference on 3D vision (3DV) (pp. 565-571). Ieee.
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).
https://www.kaggle.com/datasets/narendrakumarsubdtce/ancient-temple-vimana-images-dataset Ancient Temple Vimana images Dataset .
https://jnnce.ac.in/TempleDataSets/NARENDRA%20Description_of_KU-UBDTCE-JNNCE_Temple_Vimana_Dataset.pdf Ancient Temple Vimana images Dataset.
Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
Ioffe, S., & Szegedy, C. (2015, June). Batch normalization: Accelerating deep network training by reducing internal covariate shift. In International conference on machine learning (pp. 448-456). pmlr.
Kohavi, R. (1995, August). A study of cross-validation and bootstrap for accuracy estimation and model selection. In Ijcai (Vol. 14, No. 2, pp. 1137-1145).
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.