Comparison of iPhone and DSLR photogrammetry methods with implications for geometric morphometric analysis
This study demonstrates that iPhone photogrammetry using the Abound app produces 3D mesh data with negligible geometric differences compared to DSLR methods, validating its use for creating accessible, multi-source datasets in geometric morphometric analyses while noting that mesh source should still be accounted for in statistical models.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Museum collections are vast repositories of life's history, holding thousands of specimens that tell the story of evolution and adaptation. For decades, scientists studying the shape of these objects, such as animal skulls, relied on expensive and time-consuming machines like CT scanners to create detailed three-dimensional models. These models allowed researchers to measure subtle differences in bone structure without ever touching the fragile original. However, the high cost and complexity of such equipment meant that only a few researchers could access these digital tools, leaving most museum collections locked away from global study. In recent years, a simpler alternative has emerged: using standard cameras to take many photographs of an object from different angles and then using computer software to stitch those images together into a 3D model. This process, known as photogrammetry, has made it possible for more scientists to build digital archives, but it has also raised a new question. As researchers begin to mix models created by different cameras and different software programs into a single study, they need to know if the differences in how the models were made are larger than the actual biological differences they are trying to measure.
A team of researchers at the University of Wyoming set out to answer this question by comparing two very different ways of creating 3D models of raccoon skulls. On one side, they used a traditional, high-end digital camera and professional software, a method that produces highly detailed but complex models. On the other side, they used a smartphone app called Abound, which guides a user to take photos with an iPhone and then automatically builds the model in the cloud. The researchers wanted to see if the models made by the phone were accurate enough to be used alongside the models made by the professional equipment. They tested twenty-nine raccoon skulls, creating a digital version of each one using both methods. To ensure a fair test, they also processed the models in three different ways: leaving them as they were originally created, simplifying them to have fewer points, and smoothing them out to make the surface texture even. This allowed the team to see if the differences between the phone and the camera would disappear or grow when the models were prepared for scientific analysis.
The results showed that the two methods produced models that were remarkably similar in their overall shape. When the researchers measured the distance between the surfaces of the paired models, they found that the largest difference was only 0.18 percent of the total size of the skull, which translates to about 0.26 millimeters. To put this in perspective, this tiny gap is roughly the width of a single human hair. While the professional camera and software created models with many more tiny points and faces, making them look rougher in some areas, the smartphone app created models that were smoother and more complete. The phone app was particularly good at filling in gaps where the camera might have struggled, such as deep crevices inside the skull, by creating a smooth, watertight surface. In contrast, the professional software sometimes left holes in these difficult areas because it refused to guess what the surface looked like when the light was poor.
When the researchers analyzed the shapes to see what caused the most variation, they found that the identity of the individual raccoon was by far the most important factor. The differences between the skulls of different raccoons explained between 85 and 93 percent of the variation in the data. The method used to create the model—whether it was the phone or the professional camera—accounted for less than 4 percent of the variation. This means that the biological differences between animals are so much larger than the tiny errors introduced by the camera or software that the two types of models can be safely mixed together in a single study. The researchers noted that while the phone models were slightly less detailed on very shiny surfaces like teeth, they were still accurate enough to detect important biological features, such as tooth wear or missing teeth.
This study confirms that the smartphone app is a valid tool for scientific research, offering a fast and accessible way to create high-quality 3D models without the need for expensive equipment. It suggests that museums and researchers can now combine data from many different sources, from professional labs to field researchers using phones, to build much larger and more inclusive datasets. While the professional method remains the choice for studies requiring extreme precision on tiny internal structures, the smartphone method provides a reliable and efficient alternative for most morphological studies. By proving that these different production pipelines can work together, the research opens the door for a new era of collaborative science where the barrier to entry is simply a camera and a willingness to share.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.