Stefan B. Williams

Luca Ebner, Gideon Billings, Stefan Williams

In this work, we address the problem of real-time dense depth estimation from monocular images for mobile underwater vehicles. We formulate a deep learning model that fuses sparse depth measurements from triangulated features to improve the depth predictions and solve the problem of scale ambiguity. To allow prior inputs of arbitrary sparsity, we apply a dense parameterization method. Our model extends recent state-of-the-art approaches to monocular image based depth estimation, using an efficient encoder-decoder backbone and modern lightweight transformer optimization stage to encode global context. The network is trained in a supervised fashion on the forward-looking underwater dataset, FLSea. Evaluation results on this dataset demonstrate significant improvement in depth prediction accuracy by the fusion of the sparse feature priors. In addition, without any retraining, our method achieves similar depth prediction accuracy on a downward looking dataset we collected with a diver operated camera rig, conducting a survey of a coral reef. The method achieves real-time performance, running at 160 FPS on a laptop GPU and 7 FPS on a single CPU core and is suitable for direct deployment on embedded systems. The implementation of this work is made publicly available at https://github.com/ebnerluca/uw_depth.

Ziting Wen, Oscar Pizarro, Stefan Williams

Training deep models with limited annotations poses a significant challenge when applied to diverse practical domains. Employing semi-supervised learning alongside the self-supervised model offers the potential to enhance label efficiency. However, this approach faces a bottleneck in reducing the need for labels. We observed that the semi-supervised model disrupts valuable information from self-supervised learning when only limited labels are available. To address this issue, this paper proposes a simple yet effective framework, active self-semi-supervised learning (AS3L). AS3L bootstraps semi-supervised models with prior pseudo-labels (PPL). These PPLs are obtained by label propagation over self-supervised features. Based on the observations the accuracy of PPL is not only affected by the quality of features but also by the selection of the labeled samples. We develop active learning and label propagation strategies to obtain accurate PPL. Consequently, our framework can significantly improve the performance of models in the case of limited annotations while demonstrating fast convergence. On the image classification tasks across four datasets, our method outperforms the baseline by an average of 5.4\%. Additionally, it achieves the same accuracy as the baseline method in about 1/3 of the training time.

Suraj Bijjahalli, Oscar Pizarro, Stefan B. Williams

Detection of artificial objects from underwater imagery gathered by Autonomous Underwater Vehicles (AUVs) is a key requirement for many subsea applications. Real-world AUV image datasets tend to be very large and unlabelled. Furthermore, such datasets are typically imbalanced, containing few instances of objects of interest, particularly when searching for unusual objects in a scene. It is therefore, difficult to fit models capable of reliably detecting these objects. Given these factors, we propose to treat artificial objects as anomalies and detect them through a semi-supervised framework based on Variational Autoencoders (VAEs). We develop a method which clusters image data in a learned low-dimensional latent space and extracts images that are likely to contain anomalous features. We also devise an anomaly score based on extracting poorly reconstructed regions of an image. We demonstrate that by applying both methods on large image datasets, human operators can be shown candidate anomalous samples with a low false positive rate to identify objects of interest. We apply our approach to real seafloor imagery gathered by an AUV and evaluate its sensitivity to the dimensionality of the latent representation used by the VAE. We evaluate the precision-recall tradeoff and demonstrate that by choosing an appropriate latent dimensionality and threshold, we are able to achieve an average precision of 0.64 on unlabelled datasets.

Heather Doig, Oscar Pizarro, Stefan B. Williams

Autonomous Underwater Vehicles (AUVs) conduct regular visual surveys of marine environments to characterise and monitor the composition and diversity of the benthos. The use of machine learning classifiers for this task is limited by the low numbers of annotations available and the many fine-grained classes involved. In addition to these challenges, there are domain shifts between image sets acquired during different AUV surveys due to changes in camera systems, imaging altitude, illumination and water column properties leading to a drop in classification performance for images from a different survey where some or all these elements may have changed. This paper proposes a framework to improve the performance of a benthic morphospecies classifier when used to classify images from a different survey compared to the training data. We adapt the SymmNet state-of-the-art Unsupervised Domain Adaptation method with an efficient bilinear pooling layer and image scaling to normalise spatial resolution, and show improved classification accuracy. We test our approach on two datasets with images from AUV surveys with different imaging payloads and locations. The results show that generic domain adaptation can be enhanced to produce a significant increase in accuracy for images from an AUV survey that differs from the training images.

Ziting Wen, Oscar Pizarro, Stefan Williams

High annotation cost for training machine learning classifiers has driven extensive research in active learning and self-supervised learning. Recent research has shown that in the context of supervised learning different active learning strategies need to be applied at various stages of the training process to ensure improved performance over the random baseline. We refer to the point where the number of available annotations changes the suitable active learning strategy as the phase transition point. In this paper, we establish that when combining active learning with self-supervised models to achieve improved performance, the phase transition point occurs earlier. It becomes challenging to determine which strategy should be used for previously unseen datasets. We argue that existing active learning algorithms are heavily influenced by the phase transition because the empirical risk over the entire active learning pool estimated by these algorithms is inaccurate and influenced by the number of labeled samples. To address this issue, we propose a novel active learning strategy, neural tangent kernel clustering-pseudo-labels (NTKCPL). It estimates empirical risk based on pseudo-labels and the model prediction with NTK approximation. We analyze the factors affecting this approximation error and design a pseudo-label clustering generation method to reduce the approximation error. We validate our method on five datasets, empirically demonstrating that it outperforms the baseline methods in most cases and is valid over a wider range of training budgets.

James Bungay, Osasenaga Emokpae, Samuel D. Relton et al.

Background: Physiological tremor is defined as an involuntary and rhythmic shaking. Tremor of the hand is a key symptom of multiple neurological diseases, and its frequency and amplitude differs according to both disease type and disease progression. In routine clinical practice, tremor frequency and amplitude are assessed by expert rating using a 0 to 4 integer scale. Such ratings are subjective and have poor inter-rater reliability. There is thus a clinical need for a practical and accurate method for objectively assessing hand tremor. Objective: to develop a proof of principle method to measure hand tremor amplitude from smartphone videos. Methods: We created a computer vision pipeline that automatically extracts salient points on the hand and produces a 1-D time series of movement due to tremor, in pixels. Using the smartphones' depth measurement, we convert this measure into real distance units. We assessed the accuracy of the method using 60 videos of simulated tremor of different amplitudes from two healthy adults. Videos were taken at distances of 50, 75 and 100 cm between hand and camera. The participants had skin tone II and VI on the Fitzpatrick scale. We compared our method to a gold-standard measurement from a slide rule. Bland-Altman methods agreement analysis indicated a bias of 0.04 cm and 95% limits of agreement from -1.27 to 1.20 cm. Furthermore, we qualitatively observed that the method was robust to differences in skin tone and limited occlusion, such as a band-aid affixed to the participant's hand. Clinical relevance: We have demonstrated how tremor amplitude can be measured from smartphone videos. In conjunction with tremor frequency, this approach could be used to help diagnose and monitor neurological diseases

Ziting Wen, Oscar Pizarro, Stefan Williams

Fine-tuning the pre-trained model with active learning holds promise for reducing annotation costs. However, this combination introduces significant computational costs, particularly with the growing scale of pre-trained models. Recent research has proposed proxy-based active learning, which pre-computes features to reduce computational costs. Yet, this approach often incurs a significant loss in active learning performance, sometimes outweighing the computational cost savings. This paper demonstrates that not all sample selection differences result in performance degradation. Furthermore, we show that suitable training methods can mitigate the decline of active learning performance caused by certain selection discrepancies. Building upon detailed analysis, we propose a novel method, aligned selection via proxy, which improves proxy-based active learning performance by updating pre-computed features and selecting a proper training method. Extensive experiments validate that our method improves the total cost of efficient active learning while maintaining computational efficiency. The code is available at \url{https://github.com/ZiTingW/asvp}.

Heather Doig, Oscar Pizarro, Jacquomo Monk et al.

One use of Autonomous Underwater Vehicles (AUVs) is the monitoring of habitats associated with threatened, endangered and protected marine species, such as the handfish of Tasmania, Australia. Seafloor imagery collected by AUVs can be used to identify individuals within their broader habitat context, but the sheer volume of imagery collected can overwhelm efforts to locate rare or cryptic individuals. Machine learning models can be used to identify the presence of a particular species in images using a trained object detector, but the lack of training examples reduces detection performance, particularly for rare species that may only have a small number of examples in the wild. In this paper, inspired by recent work in few-shot learning, images and annotations of common marine species are exploited to enhance the ability of the detector to identify rare and cryptic species. Annotated images of six common marine species are used in two ways. Firstly, the common species are used in a pre-training step to allow the backbone to create rich features for marine species. Secondly, a copy-paste operation is used with the common species images to augment the training data. While annotations for more common marine species are available in public datasets, they are often in point format, which is unsuitable for training an object detector. A popular semantic segmentation model efficiently generates bounding box annotations for training from the available point annotations. Our proposed framework is applied to AUV images of handfish, increasing average precision by up to 48\% compared to baseline object detection training. This approach can be applied to other objects with low numbers of annotations and promises to increase the ability to actively monitor threatened, endangered and protected species.

Wilhelm Johan Marais, Stefan B. Williams, Oscar Pizarro

This paper presents methods for finding optimal configurations and actuator forces/torques to maximise contact wrenches in a desired direction for Underwater Vehicles Manipulator Systems (UVMS). The wrench maximisation problem is formulated as a linear programming problem, and the optimal configuration is solved as a bi-level optimisation in the parameterised redundancy space. We additionally consider the cases of one or more manipulators with multiple contact forces, maximising wrench capability while tracking a trajectory, and generating large wrench impulses using dynamic motions. We look at the specific cases of maximising force to lift a heavy load, and maximising torque during a valve turning operation. Extensive experimental results are presented using an underwater robotic platform equipped with a 4DOF manipulator, and show significant increases in wrench capability compared to existing methods for UVMS.

Takaki Yamada, Adam Prügel-Bennett, Stefan B. Williams et al.

This paper describes Georeference Contrastive Learning of visual Representation (GeoCLR) for efficient training of deep-learning Convolutional Neural Networks (CNNs). The method leverages georeference information by generating a similar image pair using images taken of nearby locations, and contrasting these with an image pair that is far apart. The underlying assumption is that images gathered within a close distance are more likely to have similar visual appearance, where this can be reasonably satisfied in seafloor robotic imaging applications where image footprints are limited to edge lengths of a few metres and are taken so that they overlap along a vehicle's trajectory, whereas seafloor substrates and habitats have patch sizes that are far larger. A key advantage of this method is that it is self-supervised and does not require any human input for CNN training. The method is computationally efficient, where results can be generated between dives during multi-day AUV missions using computational resources that would be accessible during most oceanic field trials. We apply GeoCLR to habitat classification on a dataset that consists of ~86k images gathered using an Autonomous Underwater Vehicle (AUV). We demonstrate how the latent representations generated by GeoCLR can be used to efficiently guide human annotation efforts, where the semi-supervised framework improves classification accuracy by an average of 10.2% compared to the state-of-the-art SimCLR using the same CNN and equivalent number of human annotations for training.

Jackson Shields, Oscar Pizarro, Stefan B. Williams

Special-purpose Autonomous Underwater Vehicles (AUVs) are utilised for benthic (seafloor) surveys, where the vehicle collects optical imagery of the seafloor. Due to the small-sensor footprint of the cameras and the vast areas to be surveyed, these AUVs can not feasibly collect full coverage imagery of areas larger than a few tens of thousands of square meters. Therefore it is necessary for AUV paths to sample the surveys areas sparsely, yet effectively. Broad-scale acoustic bathymetric data is readily available over large areas, and is often a useful prior of seafloor cover. As such, prior bathymetry can be used to guide AUV data collection. This research proposes methods for planning initial AUV surveys that efficiently explore a feature space representation of the bathymetry, in order to sample from a diverse set of bathymetric terrain. This will enable the AUV to visit areas that likely contain unique habitats and are representative of the entire survey site. We propose several information gathering planners that utilise a feature space exploration reward, to plan freeform paths or to optimise the placement of a survey template. The suitability of these methods to plan AUV surveys is evaluated based on the coverage of the feature space and also the ability to visit all classes of benthic habitat on the initial dive. Informative planners based on Rapidly-expanding Random Trees (RRT) and Monte-Carlo Tree Search (MCTS) were found to be the most effective. This is a valuable tool for AUV surveys as it increases the utility of initial dives. It also delivers a comprehensive training set to learn a relationship between acoustic bathymetry and visually-derived seafloor classifications.

Jackson Shields, Oscar Pizarro, Stefan B. Williams

Autonomous Underwater Vehicles (AUVs) are increasingly being used to support scientific research and monitoring studies. One such application is in benthic habitat mapping where these vehicles collect seafloor imagery that complements broadscale bathymetric data collected using sonar. Using these two data sources, the relationship between remotely-sensed acoustic data and the sampled imagery can be learned, creating a habitat model. As the areas to be mapped are often very large and AUV systems collecting seafloor imagery can only sample from a small portion of the survey area, the information gathered should be maximised for each deployment. This paper illustrates how the habitat models themselves can be used to plan more efficient AUV surveys by identifying where to collect further samples in order to most improve the habitat model. A Bayesian neural network is used to predict visually-derived habitat classes when given broad-scale bathymetric data. This network can also estimate the uncertainty associated with a prediction, which can be deconstructed into its aleatoric (data) and epistemic (model) components. We demonstrate how these structured uncertainty estimates can be utilised to improve the model with fewer samples. Such adaptive approaches to benthic surveys have the potential to reduce costs by prioritizing further sampling efforts. We illustrate the effectiveness of the proposed approach using data collected by an AUV on offshore reefs in Tasmania, Australia.

Eric L. Ferguson, Stefan B. Williams, Craig T. Jin

When a motor-powered vessel travels past a fixed hydrophone in a multipath environment, a Lloyd's mirror constructive/destructive interference pattern is observed in the output spectrogram. The power cepstrum detects the periodic structure of the Lloyd's mirror pattern by generating a sequence of pulses (rahmonics) located at the fundamental quefrency (periodic time) and its multiples. This sequence is referred to here as the `rahmonic component' of the power cepstrum. The fundamental quefrency, which is the reciprocal of the frequency difference between adjacent interference fringes, equates to the multipath time delay. The other component of the power cepstrum is the non-rahmonic (extraneous) component, which combines with the rahmonic component to form the (total) power cepstrum. A data processing technique, termed `cepstrum subtraction', is described. This technique suppresses the extraneous component of the power cepstrum, leaving the rahmonic component that contains the desired multipath time delay information. This technique is applied to real acoustic recordings of motor-vessel transits in a shallow water environment, where the broadband noise radiated by the vessel arrives at the hydrophone via a direct ray path and a time-delayed multipath. The results show that cepstrum subtraction improves multipath time delay estimation by a factor of two for the at-sea experiment. keywords - time delay estimation, underwater acoustics, cepstrum, source localization, autocorrelation

Eric L. Ferguson, Stefan B. Williams, Craig T. Jin

The propagation of sound in a shallow water environment is characterized by boundary reflections from the sea surface and sea floor. These reflections result in multiple (indirect) sound propagation paths, which can degrade the performance of passive sound source localization methods. This paper proposes the use of convolutional neural networks (CNNs) for the localization of sources of broadband acoustic radiated noise (such as motor vessels) in shallow water multipath environments. It is shown that CNNs operating on cepstrogram and generalized cross-correlogram inputs are able to more reliably estimate the instantaneous range and bearing of transiting motor vessels when the source localization performance of conventional passive ranging methods is degraded. The ensuing improvement in source localization performance is demonstrated using real data collected during an at-sea experiment.

Eric L. Ferguson, Rishi Ramakrishnan, Stefan B. Williams et al.

A cost effective approach to remote monitoring of protected areas such as marine reserves and restricted naval waters is to use passive sonar to detect, classify, localize, and track marine vessel activity (including small boats and autonomous underwater vehicles). Cepstral analysis of underwater acoustic data enables the time delay between the direct path arrival and the first multipath arrival to be measured, which in turn enables estimation of the instantaneous range of the source (a small boat). However, this conventional method is limited to ranges where the Lloyd's mirror effect (interference pattern formed between the direct and first multipath arrivals) is discernible. This paper proposes the use of convolutional neural networks (CNNs) for the joint detection and ranging of broadband acoustic noise sources such as marine vessels in conjunction with a data augmentation approach for improving network performance in varied signal-to-noise ratio (SNR) situations. Performance is compared with a conventional passive sonar ranging method for monitoring marine vessel activity using real data from a single hydrophone mounted above the sea floor. It is shown that CNNs operating on cepstrum data are able to detect the presence and estimate the range of transiting vessels at greater distances than the conventional method.

Troy Wilson, Stefan B. Williams

This paper describes the design, implementation and testing of a suite of algorithms to enable depth constrained autonomous bathymetric (underwater topography) mapping by an Autonomous Surface Vessel (ASV). Given a target depth and a bounding polygon, the ASV will find and follow the intersection of the bounding polygon and the depth contour as modeled online with a Gaussian Process (GP). This intersection, once mapped, will then be used as a boundary within which a path will be planned for coverage to build a map of the Bathymetry. Methods for sequential updates to GP's are described allowing online fitting, prediction and hyper-parameter optimisation on a small embedded PC. New algorithms are introduced for the partitioning of convex polygons to allow efficient path planning for coverage. These algorithms are tested both in simulation and in the field with a small twin hull differential thrust vessel built for the task.