In this study, we used convolutional neural networks (CNNs) to try and automate the classification of polish traces on stone tools — basically, the shiny or smooth spots that develop on tools from use-wear (e.g., from cutting wood, hide, or bone). We set up experiments with known contact materials (wood, hide, bone) and different use-intensity durations. We also experimented with different imaging parameters (like microscope objective magnification, patch size of image segments) and compared different model architectures (custom CNN vs pre-trained models like ResNet50). We looked into the following questions, * Whether a CNN can reliably tell the difference between polish made by contact with bone, hide or wood. * Whether the duration of use (short term vs long term) affects how easily the model classifies polish * How the imaging setup and model design affect performance (does using a smaller patch of the tool surface versus a larger one matter? Which objective magnification works best? Does a model trained from scratch outperform one that is pre-trained?). We found that CNN approach can classify polish from bone and hide quite well, but struggles more with wood. And also, the models were also able to distinguish polish from short-duration use versus longer use. We also emphasise that to really make this kind of approach robust, the datasets need to be much larger and more varied (different raw materials, contact materials, use contexts) and that imaging/analysis workflows should be optimised. It shows that deep learning has real promise in helping archaeologists analyse use-wear on tools in a more automated, standardised way, which could reduce human bias and increase repeatability. But it also cautions that there are still big challenges, small datasets, limited raw-material/usage variation, and model/generalisation issues mean we’re not yet at the point of “plug-and-play” ML for use-wear. For other researchers (working on archaeological objects, use-wear, materials) this paper suggests that combining controlled experiments + imaging + ML could open new lines of investigation, but we all need to pay attention to dataset design, imaging setup, and model choice.
In this paper we looked at how early hominins (tool‐making humans) at the site complex Melka Wakena in the Ethiopian highlands chose different stone materials for making percussive tools (like hammerstones). We gathered stone samples of different rock types found in that area. We measured their physical and engineering properties to see how the raw materials might differ in performance. We found measurable differences in the rock types, harder and denser rocks tended to show less volume loss (for example less damage) under the same experimental stress. For example, one rock (glassy ignimbrite) showed very low visible damage at macroscopic scale compared with others. It gives us concrete evidence that early humans weren’t just randomly picking stones; they may have been making informed decisions about which rocks to use for tools based on their likely durability and suitability for the task. This adds to our understanding of cognitive and technological behaviours in deep prehistory. Also, by establishing a baseline of raw material properties, we can later compare actual archaeological tools and see how performance and raw material choice relate.
In this experiment we tested how different stone raw materials perform as tools—specifically looking at four types, flint, dacite, quartzite and obsidian. We cut into pinewood under identical conditions and measured two key things, how deep the tool edge cuts (effectiveness) and how much the edge wore down (durability). We also measured material‐properties like hardness and grain size to see how those link to performance. What we looked at, Does tool edge efficiency vary with raw material type? (Our null hypothesis was that it doesn’t.) Which materials cut the deepest and also resist wear the best? How do traditional categories of high quality vs low quality raw materials hold up when tested quantitatively under use? We found that the different raw materials behaved very differently, flint showed the highest overall performance (deep cuts and moderate wear). Quartzite, though often considered low quality performed surprisingly well once the initial edge fragmentation settled. Obsidian was brittle and wore fast; dacite had good early performance but then plateaued. Additionally, the classic fine‐grain vs coarse‐grain raw material classification (flint is fine, quartzite is coarse) is too simplistic, some coarse‐grained materials (like quartzite here) can be efficient tool materials under certain conditions. For archaeologists this study gives us experimental data that supports the idea that raw material choice was not random. People in the past likely made technological decisions based on performance of the tool‐materials, not just availability. It also suggests that when you see certain raw materials in the archaeological record, you might infer something about the tasks they were used for, the mobility or procurement strategies of tool‐makers, or even maintenance and resharpening practices.