Personalised dietary advice has become increasingly popular, currently however most approaches are based on an individual\'s genetic and phenotypic profile whilst largely ignoring other determinants such as socio economic and cognitive variables. This paper provides novel insights by testing the effectiveness of personalised healthy eating advice concurrently tailored to an individual\'s socio-demographic group, cognitive characteristics, and sensory preferences. We first used existing data to build a synthetic dataset based on information from 3654 households (Study 1a), and then developed a cluster model to identify individuals characterised by similar socio-demographic, cognitive, and sensory aspects (Study 1b). Finally, in Study 2 we used the characteristics of 8 clusters to build 8 separate personalised food choice advice and assess their ability to motivate the increased consumption of fruit and vegetables and decreased intakes of saturated fat and sugar. We presented 218 participants with either generic UK Government \"EatWell\" advice, advice that was tailored to their allocated cluster (matched personalised), or advice tailored to a different cluster (unmatched personalised). Results showed that, when compared to generic advice, participants that received matched personalised advice were significantly more likely to indicate they would change their diet. Participants were similarly motivated to increase vegetable consumption and decrease saturated fat intake when they received unmatched personalised advice, potentially highlighting the power of providing alternative food choices. Overall, this study demonstrated that the power of personalizing food choice advice, based on a combination of individual characteristics, can be more effective than current approaches in motivating dietary change. Our study also emphasizes the viability of addressing population health through automatically delivered web-based personalised advice.

Optical tracking of head pose via fiducial markers has been proven to enable effective correction of motion artifacts in the brain during magnetic resonance imaging but remains difficult to implement in the clinic due to lengthy calibration and set up times. Advances in deep learning for markerless head pose estimation have yet to be applied to this problem because of the sub-millimetre spatial resolution required for motion correction. In the present work, two optical tracking systems are described for the development and training of a neural network: one marker-based system (a testing platform for measuring ground truth head pose) with high tracking fidelity to act as the training labels, and one markerless deep-learning-based system using images of the markerless head as input to the network. The markerless system has the potential to overcome issues of marker occlusion, insufficient rigid attachment of the marker, lengthy calibration times, and unequal performance across degrees of freedom (DOF), all of which hamper the adoption of marker-based solutions in the clinic. Detail is provided on the development of a custom moiré-enhanced fiducial marker for use as ground truth and on the calibration procedure for both optical tracking systems. Additionally, the development of a synthetic head pose dataset is described for the proof of concept and initial pre-training of a simple convolutional neural network. Results indicate that the ground truth system has been sufficiently calibrated and can track head pose with an error of <1 mm and <1°. Tracking data of a healthy, adult participant are shown. Pre-training results show that the average root-mean-squared error across the 6 DOF is 0.13 and 0.36 (mm or degrees) on a head model included and excluded from the training dataset, respectively. Overall, this work indicates excellent feasibility of the deep-learning-based approach and will enable future work in training and testing on a real dataset in the MRI environment.

Disease prediction is greatly challenged by the scarcity of datasets and privacy concerns associated with real medical data. An approach that stands out to circumvent this hurdle is the use of synthetic data generated using Generative Adversarial Networks (GANs). GANs can increase data volume while generating synthetic datasets that have no direct link to personal information. This study pioneers the use of GANs to create synthetic datasets and datasets augmented using traditional augmentation techniques for our binary classification task. The primary aim of this research was to evaluate the performance of our novel Conditional Deep Convolutional Neural Network (C-DCNN) model in classifying brain tumors by leveraging these augmented and synthetic datasets. We utilized advanced GAN models, including Conditional Deep Convolutional Generative Adversarial Network (DCGAN), to produce synthetic data that retained essential characteristics of the original datasets while ensuring privacy protection. Our C-DCNN model was trained on both augmented and synthetic datasets, and its performance was benchmarked against state-of-the-art models such as ResNet50, VGG16, VGG19, and InceptionV3. The evaluation metrics demonstrated that our C-DCNN model achieved accuracy, precision, recall, and F1 scores of 99% on both synthetic and augmented images, outperforming the comparative models. The findings of this study highlight the potential of using GAN-generated synthetic data in enhancing the training of machine learning models for medical image classification, particularly in scenarios with limited data available. This approach not only improves model accuracy but also addresses privacy concerns, making it a viable solution for real-world clinical applications in disease prediction and diagnosis.

Shadow, a natural phenomenon resulting from the absence of light, plays a pivotal role in agriculture, particularly in processes such as photosynthesis in plants. Despite the availability of generic shadow datasets, many suffer from annotation errors and lack detailed representations of agricultural shadows with possible human activity inside, excluding those derived from satellite or drone views. In this paper, we present an evaluation of a synthetically generated top-down shadow segmentation dataset characterized by photorealistic rendering and accurate shadow masks. We aim to determine its efficacy compared to real-world datasets and assess how factors such as annotation quality and image domain influence neural network model training. To establish a baseline, we trained numerous baseline architectures and subsequently explored transfer learning using various freely available shadow datasets. We further evaluated the out-of-domain performance compared to the training set of other shadow datasets. Our findings suggest that AgroSegNet demonstrates competitive performance and is effective for transfer learning, particularly in domains similar to agriculture.

Several studies in computer vision have examined specular removal, which is crucial for object detection and recognition. This research has traditionally been divided into two tasks: specular highlight removal, which focuses on removing specular highlights on object surfaces, and reflection removal, which deals with specular reflections occurring on glass surfaces. In reality, however, both types of specular effects often coexist, making it a fundamental challenge that has not been adequately addressed. Recognizing the necessity of integrating specular components handled in both tasks, we constructed a specular-light (S-Light) DB for training single-image-based deep learning models. Moreover, considering the absence of benchmark datasets for quantitative evaluation, the multi-scale normalized cross correlation (MS-NCC) metric, which considers the correlation between specular and diffuse components, was introduced to assess the learning outcomes.

Vehicle detection is a very important aspect of computer vision application to aerial and satellite imagery, facilitating activities such as instance counting, velocity estimation, traffic predictions, etc. The feasibility of accurate vehicle detection often depends on limited training datasets, requiring a lot of manual work in collection and annotation tasks. Furthermore, there are no known publicly available datasets. Our aim was to construct a pipeline for synthetic dataset generation from aerial imagery and 3D models in Blender software. The dataset generation pipeline consists of seven steps and results in a wished number of images with bounding boxes in YOLO and coco formats. This synthetic dataset has been produced following the steps described in this pipeline. It consists of 5000 2048 × 2048 images with cars inserted into the roads and highways at the images without cars from all over the world. We believe that this dataset and the respective pipeline might be of great importance for vehicle detection, facilitating the customizability of the models to specific needs and context.

Visual localization refers to the process of determining an observer\'s pose by analyzing the spatial relationships between a query image and a pre-existing set of images. In this procedure, matched visual features between images are identified and utilized for pose estimation; consequently, the accuracy of the estimation heavily relies on the precision of feature matching. Incorrect feature matchings, such as those between different objects and/or different points within an object in an image, should thus be avoided. In this paper, our initial evaluation focused on gauging the reliability of each object class within image datasets concerning pose estimation accuracy. This assessment revealed the building class to be reliable, while humans exhibited unreliability across diverse locations. The subsequent study delved deeper into the degradation of pose estimation accuracy by artificially increasing the proportion of the unreliable object-humans. The findings revealed a noteworthy decline started when the average proportion of the humans in the images exceeded 20%. We discuss the results and implications for dataset construction for visual localization.

This study demonstrates how to generate a three-dimensional (3D) body model through a small number of images and derive body values similar to the actual values using generated 3D body data. In this study, a 3D body model that can be used for body type diagnosis was developed using two full-body pictures of the front and side taken with a mobile phone. For data training, 400 3D body datasets (male: 200, female: 200) provided by Size Korea were used, and four models, i.e., 3D recurrent reconstruction neural network, point cloud generative adversarial network, skinned multi-person linear model, and pixel-aligned impact function for high-resolution 3D human digitization, were used. The models proposed in this study were analyzed and compared. A total of 10 men and women were analyzed, and their corresponding 3D models were verified by comparing 3D body data derived from 2D image inputs with those obtained using a body scanner. The model was verified through the difference between 3D data derived from the 2D image and those derived using an actual body scanner. Unlike the 3D generation models that could not be used to derive the body values in this study, the proposed model was successfully used to derive various body values, indicating that this model can be implemented to identify various body types and monitor obesity in the future.

Due to the difficulty in generating a 6-Degree-of-Freedom (6-DoF) object pose estimation dataset, and the existence of domain gaps between synthetic and real data, existing pose estimation methods face challenges in improving accuracy and generalization. This paper proposes a methodology that employs higher quality datasets and deep learning-based methods to reduce the problem of domain gaps between synthetic and real data and enhance the accuracy of pose estimation. The high-quality dataset is obtained from Blenderproc and it is innovatively processed using bilateral filtering to reduce the gap. A novel attention-based mask region-based convolutional neural network (R-CNN) is proposed to reduce the computation cost and improve the model detection accuracy. Meanwhile, an improved feature pyramidal network (iFPN) is achieved by adding a layer of bottom-up paths to extract the internalization of features of the underlying layer. Consequently, a novel convolutional block attention module-convolutional denoising autoencoder (CBAM-CDAE) network is proposed by presenting channel attention and spatial attention mechanisms to improve the ability of AE to extract images\' features. Finally, an accurate 6-DoF object pose is obtained through pose refinement. The proposed approach is compared to other models using the T-LESS and LineMOD datasets. Comparison results demonstrate the proposed approach outperforms the other estimation models.

Pose estimation of C. elegans in image sequences is challenging and even more difficult in low-resolution images. Problems range from occlusions, loss of worm identity, and overlaps to aggregations that are too complex or difficult to resolve, even for the human eye. Neural networks, on the other hand, have shown good results in both low-resolution and high-resolution images. However, training in a neural network model requires a very large and balanced dataset, which is sometimes impossible or too expensive to obtain. In this article, a novel method for predicting C. elegans poses in cases of multi-worm aggregation and aggregation with noise is proposed. To solve this problem we use an improved U-Net model capable of obtaining images of the next aggregated worm posture. This neural network model was trained/validated using a custom-generated dataset with a synthetic image simulator. Subsequently, tested with a dataset of real images. The results obtained were greater than 75% in precision and 0.65 with Intersection over Union (IoU) values.