See below details of projects involving David Labonte.
|David Labonte||3D shape analysis via flow fields||Lab based||Biomechanics and mechanobiology,Biomedical sensing diagnostics and imaging,Computational and theoretical modelling||Any two objects may differ in size and in shape. Size differences are typically obvious, but shape differences can be subtle, and challenging to quantify, because typical methods involve 2D measures such as characteristic lengths. Recent advances in computational methods have enabled detailed quantitative analyses on complex 3D shapes. In this project, you will learn and deploy one approach, based on a shape atlas. This approach is applicable to any set of 3D objects, and thus has substantial generality.Generating a shape atlas requires to estimate an anatomical model (i.e. template) as the mean of a set of input shapes. Subsequently, shape variation is quantified via the calculation of the deformation required to mold this mean shape onto a shape representing an individual from the population of relevant shapes. Mathematically, this approach represents deformation between shapes as the diffeomorphic transformation of flow fields. By means of a gradient descent optimisation scheme, the method is able to produce a statistical atlas of the population of shapes, so allowing both quantitative and detailed visual comparison of shape differences across objects.You will use this approach to investigate differences in shape across workers of a colony formed by social insects - leaf-cutter ants. Workers of this important pest species differ in weight by more than two orders of magnitude, and it is commonly speculated that this size-differences goes hand in hand with a task specialisation, leading to the definition of "worker castes". However, it remains entirely unclear whether such shape differences exist. The required workflow is established in the laboratory, and µCT scans of ant workers of different sizes are available. Your main task will be to process these scans using CT-segmentation software and Blender, to then deploy the developed workflow, and produce the shape Atlas, which will enable you to answer if ant workers of different castes differ only in size or also in shape. Recommended literature:Toussaint, N., Redhead, Y., Vidal-García, M., Lo Vercio, L., Liu, W., Fisher, E. M., ... & Green, J. B. (2021). A landmark-free morphometrics pipeline for high-resolution phenotyping: application to a mouse model of Down syndrome. Development, 148(18), dev188631.|
|David Labonte||AntGate: a colony door which selects for ant size||Lab based||Biomechanics and mechanobiology,Computational and theoretical modelling,Neurotechnology and robotics||Many experiments on animals require identification of different individuals—often a difficult task for the human eye. Traditional methods are laborious, involving marking each individual and closely following them over time, necessitating the development of automated methods. The goal of AntGate is to develop a door that opens only for permitted insects, improving the efficiency and reliability of research, while allowing exploration of questions on individual learning and behaviour.Using computer vision models developed in the lab, the mass of an insect will be extracted from a camera located next to the gate, and the gate will be activated if the insect is of appropriate size. The camera will also be able recognize tags, using existing technology , and trigger the gate if the insect is on the permitted list . The core challenge of this project will be engineering a tunnel and gate system seamlessly integrated with the camera input that allows only one insect through at a time. The main insect used to develop this gate will be ants, as their large colony sizes increase the need for automated tools. In this project, you will gain experience in machine learning techniques, camera control, and the building and development of practical tools for research. The system will be used to investigate how size and experience impacts ant locomotion, energetics, and behaviour, contributing to important advances in biomechanics and complex systems. Recommended Literature: Crall, J.D., Gravish, N., Mountcastle, A.M. and Combes, S.A., 2015. BEEtag: a low-cost, image-based tracking system for the study of animal behavior and locomotion. PloS one, 10(9), p.e0136487. Robinson, E.J., Feinerman, O. and Franks, N.R., 2012. Experience, corpulence and decision making in ant foraging. Journal of Experimental Biology, 215(15), pp.2653-2659.|
|David Labonte||Design of an Actively Powered Omni Directional Insect Treadmill||Lab based||Biomechanics and mechanobiology,Computational and theoretical modelling||Insects are the undisputed champions of legged locomotion, having mastered walking, running, and climbing on virtually any surface, often on steep inclines or even upside down. Therefore, insects have become the inspiration for the design of hexapod robots. To understand which adaptations allow insects to perform these incredible feats, we can use deep learning-based markerless pose estimation to study their locomotion from video recordings. However, this process requires not only intricate camera setups and extensively trained machine learning models but also large numbers of recorded gait cycles, which are both difficult and stressful to obtain - not only for the experimenter but especially the studied animal. In the past, conventional treadmills have been miniaturised and used for these applications, which come with the caveat of requiring the animal to walk a straight line and with external stimulation, potentially provoking flight responses instead of a natural gait. To this end, this project aims to design and build an actively powered omnidirectional insect treadmill that allows for the automated recording of freely walking insects.The goal is for the setup to automatically track the insect walking on the treadmill in real-time and control the motor speeds accordingly so the insect can move in any direction at will while being kept in the centre of the surrounding recording setup. As we aim to investigate the locomotion of a broad range of species walking on different substrates, the belts of the treadmill will need to be interchangeable to enable the use of various materials.As the project encompasses designing and building an omnidirectional treadmill as well as implementing its control loop, experience in Computer-Aided Design, rapid prototyping, and programming are required.For some design inspiration / suggested literature of large omnidirectional treadmillsPyo, S., Lee, H. and Yoon, J. (2021) ‘Development of a Novel Omnidirectional Treadmill-Based Locomotion Interface Device with Running Capability’, Applied Sciences, 11(9), p. 4223. doi: 10.3390/app11094223.Souman, J. L., Giordano, P. R., Schwaiger, M., Frissen, I., Thümmel, T., Ulbrich, H., De Luca, A., Bülthoff, H. H. and Ernst, M. O. (2011) ‘CyberWalk: Enabling unconstrained omnidirectional walking through virtual environments’, ACM Transactions on Applied Perception, 8(4). doi: 10.1145/2043603.2043607.|
|David Labonte||Going out on a limb: How do insects actively lose their limbs||Lab based||Biomechanics and mechanobiology,Computational and theoretical modelling||Many arthropods have evolved the ability to lose their limbs as a defensive strategy against predators. Although this strategy – called autotomy – has been extensively studied in reptiles, relatively little is known about the underlying mechanisms for autotomy in insects. In this project, you will be responsible for building a set-up to record this process in 3D and to analyse the kinematics of autotomy using several different species of insects. More specifically, you will adapt an existing design that can accommodate multiple cameras to capture the movements of the insect as it autotomises, and then use DeepLabCut to reconstruct these movements in 3D for analysis. A background in robotics or electronics is required, and an introductory-level understanding of machine learning is desirable.|
|David Labonte||Markerless pose estimation to study the locomotion of load-carrying leaf cutter ants||Lab based||Biomechanics and mechanobiology||Imagine, instead of driving your car to the nearest supermarket, you would have to carry your cars weight in groceries over your head while doing parkour during rush hour in the middle of a crowded city for twelve hours every day What sounds insane to a human is the daily life of a leaf-cutter ant. It is widely known that ants are capable of carrying loads greater than twenty times of their own body weight when transporting food back to their colony. As they live in symbiosis with a fungus they grow, the colonies survival and growth depend on the workers ability to harvest substantial amounts of plant material to feed the fungus. What enables these tiny creatures to move freely under the weight of the cut leaf fragments? How does the additional load on their joints influence their locomotion?These questions are to be investigated in this project. Instead of relying on manual evaluation of video data, you will train a deep neural network architecture based on DeepLabCut to perform pose estimation of ants, carrying different loads. This approach enables us to automate the extraction of tracking data when comparing the influence of load on workers of different sizes and potentially different species. You will also be involved in the design of a multi-camera setup to record individual workers from various angles synchronously, in order to create 3D reconstructions of the recorded gait cycles.Due to the proposed methodology, prior experience in machine learning and computer vision, as well as mechanics are required for this project. You will gain insights into the use and implementation of deep neural networks, creation and challenges of labelled training data sets, 3D reconstruction of tracking data, and the biomechanics of insect locomotion.For more info on the group: http://evo-biomech.ic.ac.ukRecommended Literature:1. Mathis, A. et al. DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nat. Neurosci. 21, 12811289 (2018).2. Zollikofer, C. P. E. Stepping patterns in ants. J. exp. Biol. 127, 119127 (1994).3. Wilson, E. O. Caste and division of labor in leaf-cutter ants I. Overall Pattern in A. sexdens. Behav. Ecol. Sociobiol. 7, 143156 (1980).4. Moll, K., Roces, F. & Federle, W. How Load-Carrying Ants Avoid Falling Over: Mechanical Stability during Foraging in Atta vollenweideri Grass-Cutting Ants. PLoS One 8, e52816 (2013).|
|David Labonte||Markerless pose-estimation to study the effects of leg loss on insect locomotion||Lab based||Biomechanics and mechanobiology,Computational and theoretical modelling||Insects regularly lose or let go of limbs, yet are still able to walk. The strategies insects use to adjust locomotion after leg loss may inspire strategies to enable robots to continue to function even when limbs are lost or damaged. In order to understand more about the adaptations deployed by insects upon leg loss, we will use deep neural network based markerless pose estimation to study insect locomotion. Our goal is to not only gather a deeper understanding of hexapod locomotion with different numbers of legs, but also to produce a versatile, robust and automated detection and tracking process for studying limb orientations in diverse climbing animals.A background in machine learning and computer vision is required for this project, as one of our main objectives is to train and evaluate an architecture based on DeepLabCut, in order to accurately estimate the limb positions of different species and investigate the transferability of the learned models. You will gain insights into the use and implementation of deep neural networks, creation and challenges of labelled training data sets, 3D reconstruction of tracking data, and the biomechanics of insect locomotion.For more information on the group: http://evo-biomech.ic.ac.uk/|
|David Labonte||The mechanical efficiency of herbivory||Lab based||Biomechanics and mechanobiology||Leafcutter ants are the principal insect pests in the Neotropics, harvesting up to 4500m² of plant matter per year. Using their sharp mandibles, they cut a variety of food sources such as leaves, fruit and flower petals. Herbivory at such a large scale requires a large amount of cutting, which in turn uses a lot of energy. The energetic cost of such an endeavour can vary based on both the size of the ant and the toughness of the plant. Here we will look at how material properties influence the amount of energy needed to cut by ants of one size. Our goal is to gain a deeper understanding of how energetic costs scale with material properties.In this project, you will investigate this question with a multi-disciplinary experimental approach. An artificial leaf system of repeatable and controllable material properties will be made out of polymers. Using an ultra-sensitive flow-through respirometry system, we will measure both resting metabolic rates, and active metabolic rates during cutting synthetic leaves. For more information on the group visit:http://evo-biomech.ic.ac.uk/|