Search

Lithic Use-Wear Analysis (LUWA) using microscopic images is an underexplored vision-for-science research area. It seeks to distinguish the worked material, which is critical for understanding archaeological artifacts, material interactions, tool functionalities, and dental records. However, this challenging task goes beyond the well-studied image classification problem for common objects. It is affected by many confounders owing to the complex wear mechanism and microscopic imaging, which makes it difficult even for human experts to identify the worked material successfully. In this paper, we investigate the following three questions on this unique vision task for the first time:(i) How well can state-of-the-art pre-trained models (like DINOv2) generalize to the rarely seen domain? (ii) How can few-shot learning be exploited for scarce microscopic images? (iii) How do the ambiguous magnification and sensing modality influence the classification accuracy? To study these, we collaborated with archaeologists and built the first open-source and the largest LUWA dataset containing 23,130 microscopic images with different magnifications and sensing modalities. Extensive experiments show that existing pre-trained models notably outperform human experts but still leave a large gap for improvements. Most importantly, the LUWA dataset provides an underexplored opportunity for vision and learning communities and complements existing image classification problems on common objects., Comment: CVPR

Currently, approximately 90% of the human population is right-handed. This handedness is due to the lateralization of the cerebral hemispheres and is controlled by brain areas involved in complex motor tasks such as making stone tools or in language. In addition to describing the evolution of laterality in humans, identifying hand preference in fossil hominids can improve our understanding of the emergence and development of complex cognitive faculties during evolution. Several fields of prehistory like palaeoanthropology or lithic analysis have already investigated handedness in fossils hominins but they face limitations due to either the incomplete or the composite state of the skeleton remains or to results replication or method application failure. Wear analysis could provide new complementary data about hand preference evolution and the development of certain complex cognitive functions using indirect evidence (use traces, micro-scars in particular) of the hand holding the stone tool during use. Controlled experiment has been carried out in order to establish a reference collection of tools used with the left and tools used with the right hand. Wear analysis was performed on this corpus using “classical” microscopic approach and geometric morphometric analysis. A machine learning algorithm, the k-NN method, was applied to verify if use traces (micro-scars) could help determine the hand holding the tool during use. The best model, based on parameters referring to invasiveness of micro-scars, was able to correctly determine the hand holding the tool with 75% accuracy.

Currently, approximately 90% of the human population is right-handed. This handedness is due to the lateralization of the cerebral hemispheres and is controlled by brain areas involved in complex motor tasks such as making stone tools or in language. In addition to describing the evolution of laterality in humans, identifying hand preference in fossil hominids can improve our understanding of the emergence and development of complex cognitive faculties during evolution. Several fields of prehistory like palaeoanthropology or lithic analysis have already investigated handedness in fossils hominins but they face limitations due to either the incomplete or the composite state of the skeleton remains or to results replication or method application failure. Wear analysis could provide new complementary data about hand preference evolution and the development of certain complex cognitive functions using indirect evidence (use traces, micro-scars in particular) of the hand holding the stone tool during use. Controlled experiment has been carried out in order to establish a reference collection of tools used with the left and tools used with the right hand. Wear analysis was performed on this corpus using “classical” microscopic approach and geometric morphometric analysis. A machine learning algorithm, the k-NN method, was applied to verify if use traces (micro-scars) could help determine the hand holding the tool during use. The best model, based on parameters referring to invasiveness of micro-scars, was able to correctly determine the hand holding the tool with 75% accuracy.

While lithic objects can potentially inform us about past adaptations and behaviors, it is important to develop a comprehensive understanding of all of the various processes that influence what we recover from the archaeological record. We argue here that many assumptions used by archaeologists to derive behavioral inferences through the definition, conceptualization, and interpretation of both individual stone artifact forms and groups of artifacts identified as assemblages do not fit squarely with what we have learned from both ethnographic sources and analyses of archaeological materials. We discuss this in terms of two fallacies. The first is the fallacy of the "desired end product" in stone artifact manufacture, which also includes our ability to recognize such end products. The second fallacy has to do with the notions that lithic assemblages represent simple accumulations of contemporary behaviors and the degree to which the composition of the depositional units we study reliably match the kinds of activities that took place. Although it is beyond the scope of this paper to offer a comprehensive set of new methodologies and theoretical perspectives to solve these problems, our goal here is to stress the importance of rethinking some of our most basic assumptions regarding the nature of lithic objects and how they become part of the archaeological record. Such a revision is needed if we want to be able to develop research questions that can be addressed with the data we have available to us.

While lithic objects can potentially inform us about past adaptations and behaviors, it is important to develop a comprehensive understanding of all of the various processes that influence what we recover from the archaeological record. We argue here that many assumptions used by archaeologists to derive behavioral inferences through the definition, conceptualization, and interpretation of both individual stone artifact forms and groups of artifacts identified as assemblages do not fit squarely with what we have learned from both ethnographic sources and analyses of archaeological materials. We discuss this in terms of two fallacies. The first is the fallacy of the "desired end product" in stone artifact manufacture, which also includes our ability to recognize such end products. The second fallacy has to do with the notions that lithic assemblages represent simple accumulations of contemporary behaviors and the degree to which the composition of the depositional units we study reliably match the kinds of activities that took place. Although it is beyond the scope of this paper to offer a comprehensive set of new methodologies and theoretical perspectives to solve these problems, our goal here is to stress the importance of rethinking some of our most basic assumptions regarding the nature of lithic objects and how they become part of the archaeological record. Such a revision is needed if we want to be able to develop research questions that can be addressed with the data we have available to us.

Owing to a thick blanket of loess and other later geological disruptions, the earliest hominins to reach Europe are hard to find. TO a handful of Possible sites, our authors add a new assemblage of lithics with a clear local context and corroborated OSL ages. Ancient humans were present in what is now Romania between 300 000 and 400 000 years ago.

Owing to a thick blanket of loess and other later geological disruptions, the earliest hominins to reach Europe are hard to find. TO a handful of Possible sites, our authors add a new assemblage of lithics with a clear local context and corroborated OSL ages. Ancient humans were present in what is now Romania between 300 000 and 400 000 years ago.

It is long been thought that many flake attributes, including both size and shape, are largely due to the morphology of a core's flaking surface, yet this has never been tested under strictly controlled conditions. Using molded glass cores with surface morphologies that highly resemble prehistoric ones, this experiment demonstrates that while core surface morphology does exhibit some influence on flake size and shape, a high degree of variation in flakes produced with the same core surface morphology shows that the effects of other independent variables, such as exterior platform angle and platform depth, have an even stronger effect. A major implication of these results is that current approaches to reconstruct prehistoric knapping strategies are overlooking significant sources of variation.

It is long been thought that many flake attributes, including both size and shape, are largely due to the morphology of a core's flaking surface, yet this has never been tested under strictly controlled conditions. Using molded glass cores with surface morphologies that highly resemble prehistoric ones, this experiment demonstrates that while core surface morphology does exhibit some influence on flake size and shape, a high degree of variation in flakes produced with the same core surface morphology shows that the effects of other independent variables, such as exterior platform angle and platform depth, have an even stronger effect. A major implication of these results is that current approaches to reconstruct prehistoric knapping strategies are overlooking significant sources of variation.

This dissertation has two goals. The first is to provide a theoretical framework for a dynamic, morphology-independent approach to stone tool assemblage systematics and to introduce a quantitative method for evaluating relationships between assemblages based on comparisons of reduction trajectories. The second goal is to use this method to assess the relevance of an oft-cited dichotomy in European Middle Paleolithic systematics, that between assemblages with a high bifacial component and those dominated by core-and-flake technologies, from the point of view of the proposed theoretical basis for classification. The data come from three major technocomplexes: the Micoquian (Germany: sites of Buhlen and Sesselfelsgrotte; Crimea, Ukraine: sites of Kabazi II, V, Starosele, Chokurcha I, Buran-Kaya III), the Mousterian of Acheulean Tradition (MTA, France: sites of Le Moustier and Pech de l'Azé I), and the Quina Mousterian (France: sites of Roc de Marsal and Jonzac). Reduction trajectories, expressed as allometric relationships between stone tool shape and size are viewed as ‘reversed’ ontogenies of stone tools, and are compared in the same way as biological ontogenetic trajectories. The methodology used in this study consists of converting shape data into elliptical Fourier (EFA) coefficients and tracking the latter's variance with size. The mathematical expressions of the change in shape as a function of size (unique to each assemblage) are linear vectors, which can be compared trigonometrically to assess overall similarity of the assemblages. The results suggest that (1) morphological variability displayed by individual Middle Paleolithic tools is very low, rendering classical typological approaches ineffective; and (2) Micoquian bifacial tool reduction trajectories are more similar to those of Quina Mousterian unifacial scrapers than they are to those of MTA bifacial tools. The implications of the hypothesis testing is that, when examined through a dynamic, historical p