Computer Science

Recent literature found that convolutional neural networks (CNN) with large filters perform well in some applications such as image semantic segmentation. Winograd transformation helps to reduce the number of multiplications in a convolution but suffers from numerical instability when the convolution filter size gets large. This work proposes a nested Winograd algorithm to iteratively decompose a large filter into a sequence of 3x3 tiles which can then be accelerated with a 3x3 Winograd algorithm. Compared with the state-of-art OLA-Winograd algorithm, the proposed algorithm reduces the multiplications by 1.41 to 3.29 times for computing 5x5 to 9x9 convolutions.

Video Retrieval is a challenging task where a text query is matched to a video or vice versa. Most of the existing approaches for addressing such a problem rely on annotations made by the users. Although simple, this approach is not always feasible in practice. In this work, we explore the application of the language-image model, CLIP, to obtain video representations without the need for said annotations. This model was explicitly trained to learn a common space where images and text can be compared. Using various techniques described in this document, we extended its application to videos, obtaining state-of-the-art results on the MSR-VTT and MSVD benchmarks.

High energy impacts at joint locations often generate highly fragmented, or comminuted, bone fractures. Current approaches for treatment require physicians to decide how to classify the fracture within a hierarchy fracture severity categories. Each category then provides a best-practice treatment scenario to obtain the best possible prognosis for the patient. This article identifies shortcomings associated with qualitative-only evaluation of fracture severity and provides new quantitative metrics that serve to address these shortcomings. We propose a system to semi-automatically extract quantitative metrics that are major indicators of fracture severity. These include: (i) fracture surface area, i.e., how much surface area was generated when the bone broke apart, and (ii) dispersion, i.e., how far the fragments have rotated and translated from their original anatomic positions. This article describes new computational tools to extract these metrics by computationally reconstructing 3D bone anatomy from CT images with a focus on tibial plafond fracture cases where difficult qualitative fracture severity cases are more prevalent. Reconstruction is accomplished within a single system that integrates several novel algorithms that identify, extract and piece-together fractured fragments in a virtual environment. Doing so provides objective quantitative measures for these fracture severity indicators. The availability of such measures provides new tools for fracture severity assessment which may lead to improved fracture treatment. This paper describes the system, the underlying algorithms and the metrics of the reconstruction results by quantitatively analyzing six clinical tibial plafond fracture cases.

Despite their impressive performance in object recognition and other tasks under standard testing conditions, deep networks often fail to generalize to out-of-distribution (o.o.d.) samples. One cause for this shortcoming is that modern architectures tend to rely on "shortcuts" - superficial features that correlate with categories without capturing deeper invariants that hold across contexts. Real-world concepts often possess a complex structure that can vary superficially across contexts, which can make the most intuitive and promising solutions in one context not generalize to others. One potential way to improve o.o.d. generalization is to assume simple solutions are unlikely to be valid across contexts and avoid them, which we refer to as the too-good-to-be-true prior. A low-capacity network (LCN) with a shallow architecture should only be able to learn surface relationships, including shortcuts. We find that LCNs can serve as shortcut detectors. Furthermore, an LCN's predictions can be used in a two-stage approach to encourage a high-capacity network (HCN) to rely on deeper invariant features that should generalize broadly. In particular, items that the LCN can master are downweighted when training the HCN. Using a modified version of the CIFAR-10 dataset in which we introduced shortcuts, we found that the two-stage LCN-HCN approach reduced reliance on shortcuts and facilitated o.o.d. generalization.

With the aim of matching a pair of instances from two different modalities, cross modality mapping has attracted growing attention in the computer vision community. Existing methods usually formulate the mapping function as the similarity measure between the pair of instance features, which are embedded to a common space. However, we observe that the relationships among the instances within a single modality (intra relations) and those between the pair of heterogeneous instances (inter relations) are insufficiently explored in previous approaches. Motivated by this, we redefine the mapping function with relational reasoning via graph modeling, and further propose a GCN-based Relational Reasoning Network (RR-Net) in which inter and intra relations are efficiently computed to universally resolve the cross modality mapping problem. Concretely, we first construct two kinds of graph, i.e., Intra Graph and Inter Graph, to respectively model intra relations and inter relations. Then RR-Net updates all the node features and edge features in an iterative manner for learning intra and inter relations simultaneously. Last, RR-Net outputs the probabilities over the edges which link a pair of heterogeneous instances to estimate the mapping results. Extensive experiments on three example tasks, i.e., image classification, social recommendation and sound recognition, clearly demonstrate the superiority and universality of our proposed model.

Micro-expression can reflect people's real emotions. Recognizing micro-expressions is difficult because they are small motions and have a short duration. As the research is deepening into micro-expression recognition, many effective features and methods have been proposed. To determine which direction of movement feature is easier for distinguishing micro-expressions, this paper selects 18 directions (including three types of horizontal, vertical and oblique movements) and proposes a new low-dimensional feature called the Histogram of Single Direction Gradient (HSDG) to study this topic. In this paper, HSDG in every direction is concatenated with LBP-TOP to obtain the LBP with Single Direction Gradient (LBP-SDG) and analyze which direction of movement feature is more discriminative for micro-expression recognition. As with some existing work, Euler Video Magnification (EVM) is employed as a preprocessing step. The experiments on the CASME II and SMIC-HS databases summarize the effective and optimal directions and demonstrate that HSDG in an optimal direction is discriminative, and the corresponding LBP-SDG achieves state-of-the-art performance using EVM.

Point clouds are essential for storage and transmission of 3D content. As they can entail significant volumes of data, point cloud compression is crucial for practical usage. Recently, point cloud geometry compression approaches based on deep neural networks have been explored. In this paper, we evaluate the ability to predict perceptual quality of typical voxel-based loss functions employed to train these networks. We find that the commonly used focal loss and weighted binary cross entropy are poorly correlated with human perception. We thus propose a perceptual loss function for 3D point clouds which outperforms existing loss functions on the ICIP2020 subjective dataset. In addition, we propose a novel truncated distance field voxel grid representation and find that it leads to sparser latent spaces and loss functions that are more correlated with perceived visual quality compared to a binary representation. The source code is available at this https URL.

Accurate and automatic segmentation of three-dimensional (3D) individual teeth from cone-beam computerized tomography (CBCT) images is a challenging problem because of the difficulty in separating an individual tooth from adjacent teeth and its surrounding alveolar bone. Thus, this paper proposes a fully automated method of identifying and segmenting 3D individual teeth from dental CBCT images. The proposed method addresses the aforementioned difficulty by developing a deep learning-based hierarchical multi-step model. First, it automatically generates upper and lower jaws panoramic images to overcome the computational complexity caused by high-dimensional data and the curse of dimensionality associated with limited training dataset. The obtained 2D panoramic images are then used to identify 2D individual teeth and capture loose- and tight- regions of interest (ROIs) of 3D individual teeth. Finally, accurate 3D individual tooth segmentation is achieved using both loose and tight ROIs. Experimental results showed that the proposed method achieved an F1-score of 93.35% for tooth identification and a Dice similarity coefficient of 94.79% for individual 3D tooth segmentation. The results demonstrate that the proposed method provides an effective clinical and practical framework for digital dentistry.

Object Detection (OD) is an important computer vision problem for industry, which can be used for quality control in the production lines, among other applications. Recently, Deep Learning (DL) methods have enabled practitioners to train OD models performing well on complex real world images. However, the adoption of these models in industry is still limited by the difficulty and the significant cost of collecting high quality training datasets. On the other hand, when applying OD to the context of production lines, CAD models of the objects to be detected are often available. In this paper, we introduce a fully automated method that uses a CAD model of an object and returns a fully trained OD model for detecting this object. To do this, we created a Blender script that generates realistic labeled datasets of images containing the object, which are then used for training the OD model. The method is validated experimentally on two practical examples, showing that this approach can generate OD models performing well on real images, while being trained only on synthetic images. The proposed method has potential to facilitate the adoption of object detection models in industry as it is easy to adapt for new objects and highly flexible. Hence, it can result in significant costs reduction, gains in productivity and improved products quality.

In fruit tree growth, pruning is an important management practice for preventing overcrowding, improving canopy access to light and promoting regrowth. Due to the slow nature of agriculture, decisions in pruning are typically made using tradition or rules of thumb rather than data-driven analysis. Many existing algorithmic, simulation-based approaches rely on high-fidelity digital captures or purely computer-generated fruit trees, and are unable to provide specific results on an orchard scale. We present a framework for suggesting pruning strategies on LiDAR-scanned commercial fruit trees using a scoring function with a focus on improving light distribution throughout the canopy. A scoring function to assess the quality of the tree shape based on its light availability and size was developed for comparative analysis between trees, and was validated against yield characteristics, demonstrating a reasonable correlation against fruit count with an R 2 score of 0.615 for avocado and 0.506 for mango. A tool was implemented for simulating pruning by algorithmically estimating which parts of a tree point cloud would be removed given specific cut points using structural analysis of the tree, validated experimentally with an average F1 score of 0.78 across 144 experiments. Finally, new pruning locations were suggested and we used the previous two stages to estimate the improvement of the tree given these suggestions. The light distribution was improved by up to 25.15\%, demonstrating a 16\% improvement over commercial pruning on a real tree, and certain cut points were discovered which improved light distribution with a smaller negative impact on tree volume. The final results suggest value in the framework as a decision making tool for commercial growers, or as a starting point for automated pruning since the entire process can be performed with little human intervention.