Example Dissertation Topics in Archaeology

Background: Research Gaps, Objectives, and Initial Reading for Literature Review

The integration of machine learning, particularly deep learning, and aerial LiDAR scanning has transformed archaeological research by providing scholars with new data about ancient landscapes. However, there are still many unanswered questions regarding the practical use of modern machine learning models on ALS data for archaeology, especially in the areas of segmentation, feature recognition, and effectively interpreting subtle elevation patterns in other locations.

Research Gaps:

A number of noteworthy difficulties have emerged when ALS data in archaeology is subjected to machine learning. Larger contextual variables in complex landscapes, like vast ponds and subtle topographical characteristics connected to low-elevation hills, are difficult for current approaches to detect and categorize [1].

Many of the deep learning models-a case in point include Convolutional Neural Networks and Vision Transformers-take such a limited context window that they tend to ignore the larger, interlinking forms or broader feature sets of a site that are necessary for the archaeological interpretation [1].

Moreover, while the capability of small local feature identification is working quite well in some models, adopting a global spatial context for how these small features interact within the larger archaeological landscape becomes impossible.

Additionally, multi-modal data, such RGB and elevation data, have been underutilized even though they have been shown to improve feature detection in blocked environments, like those under trees. Similarly, hybrid models like HybViT, which combines CNNs with ViTs, show great potential, but further study is needed to improve their integration and hence their performance on large-scale archeological datasets.

Suggested Research Objectives:

1. Increase segmentation accuracy by enhancing the ability of models to detect local features and large-scale structures in ALS data.
2. Design and study multi-modal models to seamlessly integrate elevation and radiometric data for any feature detection, especially in difficult environments such as dense forests.
3. Apply architecture optimization to model types including hybrid ViTs and hierarchical transformers for enhanced discrimination of both fine-grained local features and larger spatial contexts.
4. Conduct tests and benchmarks in real-world archaeological sites to enforce the robustness and scalability of the models.

Reading Suggestions for Initial Review

Key readings for understanding these gaps and formulating solutions include:

Evans, D., Fletcher, R. J., Pottier, C., et al. (2013). Uncovering archaeological landscapes at Angkor using LiDAR.

Ronneberger, O., Fischer, P., & Brox, T. (2015). U-Net: Convolutional networks for biomedical image segmentation.

Liu, Z., Lin, Y., Cao, Y., et al. (2021). SWIN transformer: Hierarchical vision transformer using shifted windows.

Pottier, C. (1999). Carte archéologique de la région d’Angkor - Zone Sud. PhD thesis, Université de la Sorbonne Nouvelle - Paris 3.

By exploring these references and building on the identified gaps, future research can aim to create more accurate, context-aware models for archaeological data interpretation.

Research Topic 1: Limited Context Window and Incomplete Feature Detection

Background Context:

Segmentation of large-scale features has found machine learning applications difficult, especially CNNs and ViTs in full spatial context capture of archaeological landscapes. This limitation makes detection difficult on wider structures or contextual features in ALS data [1].

Research Questions:

1. How can model context windows be extended to improve the detection of large archaeological features?
2. What hybrid architectures could better integrate local and global context for archaeological landscape interpretation?
3. How can hierarchical model structures improve feature detection across different scales in ALS data?

Potential Implications:

1. Extended context windows may lead to more accurate and comprehensive feature detection in complex archaeological landscapes, facilitating better site discovery and analysis.
2. Improvement in model coupling of local with global contexts would result in a more capable detection of large-scale features and a much taller view of ancient sites.

Reading Suggestions:

• Guyot, A., Lennon, M., Lorho, T., & Hubert-Moy, L. (2021). Combined detection and segmentation of archaeological structures from LiDAR data using a deep learning approach. Journal of Computer Applications in Archaeology, 4(1), 15-30. https://doi.org/10.5334/jcaa.105
• Verschoof-van der Vaart, W., Lambers, K., Kowalczyk, W., & Bourgeois, Q. P. J. (2020). Combining deep learning and location-based ranking for large-scale archaeological prospection of LiDAR data from the Netherlands. ISPRS International Journal of Geo-Information, 9(3), 171. https://doi.org/10.3390/ijgi9030171
• Wang, W., Xie, E., Li, X., Fan, D.-P., Song, K., Liang, D., Lu, T., Luo, P., & Shao, L. (2021). Pyramid Vision Transformer: A versatile backbone for dense prediction without convolutions. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 572-581). https://doi.org/10.1109/ICCV48922.2021.00066

Research Topic 2: Inadequate Integration of Multi-Modal Data

Background Context:

Modern models, however, still don't seem to make good synergy out of multi-modal data, such as RGB images combined with elevation data, which is quite critical for detecting features hidden by vegetation or smooth terrain changes. The significance of elevation data, as contextual information, can help to improve model accuracy in challenging environments [1].

Research Questions:

1. How can multi-modal models be optimized for improved feature detection in archaeological ALS data?
2. What methods can be used to integrate elevation data effectively with radiometric information (RGB)?
3. How does multi-modal data impact feature detection accuracy in dense vegetation areas?

Potential Implications:

1. Integrating multi-modal data could enhance detection in challenging environments like dense vegetation, leading to the discovery of features previously missed by models using only a single data source.
2. Improved integration of elevation data could provide more context, enabling models to differentiate between subtle elevation changes and large-scale features, thus improving feature accuracy.

Reading Suggestions:

• Chase, A. F., Reese-Taylor, K., Fernandez-Diaz, J. C., & Chase, D. Z. (2012). Geospatial revolution and remote sensing LiDAR in Mesoamerican archaeology.Proceedings of the National Academy of Sciences, 109(20), 7683-7689. https://doi.org/10.1073/pnas.1200426109
• Banasiak, P. Z., Berezowski, P. L., Zapłata, R., Mielcarek, M., Duraj, K., & Sterenczak, K. (2022). Semantic segmentation (U-Net) of archaeological features in airborne laser scanning—example of the Białowieza forest. Remote Sensing, 14(12), 3020. https://doi.org/10.3390/rs14123020
• Schindling, J., & Gibbes, C. (2014). LiDAR as a tool for archaeological research: A case study. Archaeological and Anthropological Sciences, 6(3), 241-252. https://doi.org/10.1007/s12520-014-0159-4

Research Topic 3: Global Context Understanding and Semantic Relationships

Background Context:

Most models imitate individual features without fully understanding the links between them in a larger context. As a result, this distinction allows for incorrectly interpretation of the archaeological environment. To capture the links between local and large-scale elements, an integrated methodology must be used [1].

Research Questions:

1. How can models be designed to better understand and represent the relationships between different features in ALS data?
2. What architectural modifications can improve semantic context understanding in archaeological feature detection?
3. How can hierarchical and multi-scale models be leveraged for improved global context awareness?

Potential Implications:

1. Appreciating relationships among features could provide accurate reconstructions of ancient landscapes and the spatial organization and cultural meanings of features therein.
2. Better semantic context in models should allow for better archaeological interpretations and perhaps better than avant-garde methods to reveal hitherto unseen patterns and spatial relationships.

Reading Suggestions:

• Evans, D., Fletcher, R. J., Pottier, C., Chevance, J.-B., Soutif, D., Tan, B. S., Im, S., Ea, D., Tin, T., Kim, S., et al. (2013). Uncovering archaeological landscapes at Angkor using LiDAR. Proceedings of the National Academy of Sciences, 110(31), 12595-12600. https://doi.org/10.1073/pnas.1313787110
• Guyot, A., Lennon, M., Lorho, T., & Hubert-Moy, L. (2021). Combined detection and segmentation of archaeological structures from LiDAR data using a deep learning approach. Journal of Computer Applications in Archaeology, 4(1), 15-30. https://doi.org/10.5334/jcaa.105
• Verschoof-van der Vaart, W., Bonhage, A., Schneider, A., Ouimet, W., & Raab, T. (2023). Automated large-scale mapping and analysis of relict charcoal hearths in Connecticut (USA) using a deep learning YOLOv4 framework. Archaeological Prospection, 30(2), 111-122. https://doi.org/10.1002/arp.1923

Conclusion

In archaeological studies, exploring new aspects helps to streamline, solve, or enhance existing problems in the academic field. However, selecting and refining a research topic that contributes to the field is a critical process that requires your complete attention.

Tutors India assists you in selecting a topic that is original and highly rich in scientific observations. Our expert research assistants make academically strong works that earn you a credit.

Contact Tutors India to avail our masters dissertation topic selection services.