Despite the rapid development of Neural Radiance Field (NeRF), the necessity of dense covers largely prohibits its wider applications. While several recent works have attempted to address this issue, they either operate with sparse views (yet still, a few of them) or on simple objects/scenes. In this work, we consider a more ambitious task: training neural radiance field, over realistically complex visual scenes, by "looking only once", i.e., using only a single view. To attain this goal, we present a Single View NeRF (SinNeRF) framework consisting of thoughtfully designed semantic and geometry regularizations. Specifically, SinNeRF constructs a semi-supervised learning process, where we introduce and propagate geometry pseudo labels and semantic pseudo labels to guide the progressive training process. Extensive experiments are conducted on complex scene benchmarks, including NeRF synthetic dataset, Local Light Field Fusion dataset, and DTU dataset. We show that even without pre-training on multi-view datasets, SinNeRF can yield photo-realistic novel-view synthesis results. Under the single image setting, SinNeRF significantly outperforms the current state-of-the-art NeRF baselines in all cases. Project page: https://vita-group.github.io/SinNeRF/
Thanks to the rapid evolvement of robotic technologies, robot mowing is emerging to liberate humans from the tedious and time-consuming landscape work. Traditionally, robot mowing is perceived as a "Coverage Path Planning" problem, with a simplification that converts non-convex obstacles into convex obstacles. Besides, the converted obstacles are commonly dilated by the robot's circumcircle for collision avoidance. However when applied to robot mowing, an obstacle in a lawn is usually non-convex, imagine a garden on the lawn, such that the mentioned obstacle processing methods would fill in some concave areas so that they are not accessible to the robot anymore and hence produce inescapable uncut areas along the lawn edge, which dulls the landscape's elegance and provokes rework. To shrink the uncut area around the lawn edge we hereby reframe the problem into a brand new problem, named the "Edge Coverage Path Planning" problem that is dedicated to path planning with the objective to cover the edge. Correspondingly, we propose two planning methods, the "big and small disk" and the "sliding chopstick" planning method to tackle the problem by leveraging image morphological processing and computational geometry skills. By validation, our proposed methods can outperform the traditional "dilation-by-circumcircle" method.
Detection of out-of-distribution (OoD) samples in the context of image classification has recently become an area of interest and active study, along with the topic of uncertainty estimation, to which it is closely related. In this paper we explore the task of OoD segmentation, which has been studied less than its classification counterpart and presents additional challenges. Segmentation is a dense prediction task for which the model's outcome for each pixel depends on its surroundings. The receptive field and the reliance on context play a role for distinguishing different classes and, correspondingly, for spotting OoD entities. We introduce MOoSe, an efficient strategy to leverage the various levels of context represented within semantic segmentation models and show that even a simple aggregation of multi-scale representations has consistently positive effects on OoD detection and uncertainty estimation.
As convolutional neural networks (CNNs) become more accurate at object recognition, their representations become more similar to the primate visual system. This finding has inspired us and other researchers to ask if the implication also runs the other way: If CNN representations become more brain-like, does the network become more accurate? Previous attempts to address this question showed very modest gains in accuracy, owing in part to limitations of the regularization method. To overcome these limitations, we developed a new neural data regularizer for CNNs that uses Deep Canonical Correlation Analysis (DCCA) to optimize the resemblance of the CNN's image representations to that of the monkey visual cortex. Using this new neural data regularizer, we see much larger performance gains in both classification accuracy and within-super-class accuracy, as compared to the previous state-of-the-art neural data regularizers. These networks are also more robust to adversarial attacks than their unregularized counterparts. Together, these results confirm that neural data regularization can push CNN performance higher, and introduces a new method that obtains a larger performance boost.
We present a method for inferring diverse 3D models of human-object interactions from images. Reasoning about how humans interact with objects in complex scenes from a single 2D image is a challenging task given ambiguities arising from the loss of information through projection. In addition, modeling 3D interactions requires the generalization ability towards diverse object categories and interaction types. We propose an action-conditioned modeling of interactions that allows us to infer diverse 3D arrangements of humans and objects without supervision on contact regions or 3D scene geometry. Our method extracts high-level commonsense knowledge from large language models (such as GPT-3), and applies them to perform 3D reasoning of human-object interactions. Our key insight is priors extracted from large language models can help in reasoning about human-object contacts from textural prompts only. We quantitatively evaluate the inferred 3D models on a large human-object interaction dataset and show how our method leads to better 3D reconstructions. We further qualitatively evaluate the effectiveness of our method on real images and demonstrate its generalizability towards interaction types and object categories.
Contemporary transfer learning-based methods to alleviate the data insufficiency in change detection (CD) are mainly based on ImageNet pre-training. Self-supervised learning (SSL) has recently been introduced to remote sensing (RS) for learning in-domain representations. Here, we propose a semantic decoupled representation learning for RS image CD. Typically, the object of interest (e.g., building) is relatively small compared to the vast background. Different from existing methods expressing an image into one representation vector that may be dominated by irrelevant land-covers, we disentangle representations of different semantic regions by leveraging the semantic mask. We additionally force the model to distinguish different semantic representations, which benefits the recognition of objects of interest in the downstream CD task. We construct a dataset of bitemporal images with semantic masks in an effortless manner for pre-training. Experiments on two CD datasets show our model outperforms ImageNet pre-training, in-domain supervised pre-training, and several recent SSL methods.
To overcome inherent hardware limitations of hyperspectral imaging systems with respect to their spatial resolution, fusion-based hyperspectral image (HSI) super-resolution is attracting increasing attention. This technique aims to fuse a low-resolution (LR) HSI and a conventional high-resolution (HR) RGB image in order to obtain an HR HSI. Recently, deep learning architectures have been used to address the HSI super-resolution problem and have achieved remarkable performance. However, they ignore the degradation model even though this model has a clear physical interpretation and may contribute to improve the performance. We address this problem by proposing a method that, on the one hand, makes use of the linear degradation model in the data-fidelity term of the objective function and, on the other hand, utilizes the output of a convolutional neural network for designing a deep prior regularizer in spectral and spatial gradient domains. Experiments show the performance improvement achieved with this strategy.
The adoption of Convolutional Neural Network (CNN) models in high-stake domains is hindered by their inability to meet society's demand for transparency in decision-making. So far, a growing number of methodologies have emerged for developing CNN models that are interpretable by design. However, such models are not capable of providing interpretations in accordance with human perception, while maintaining competent performance. In this paper, we tackle these challenges with a novel, general framework for instantiating inherently interpretable CNN models, named E Pluribus Unum Interpretable CNN (EPU-CNN). An EPU-CNN model consists of CNN sub-networks, each of which receives a different representation of an input image expressing a perceptual feature, such as color or texture. The output of an EPU-CNN model consists of the classification prediction and its interpretation, in terms of relative contributions of perceptual features in different regions of the input image. EPU-CNN models have been extensively evaluated on various publicly available datasets, as well as a contributed benchmark dataset. Medical datasets are used to demonstrate the applicability of EPU-CNN for risk-sensitive decisions in medicine. The experimental results indicate that EPU-CNN models can achieve a comparable or better classification performance than other CNN architectures while providing humanly perceivable interpretations.
Given an input image, and nothing else, our method returns the bounding boxes of objects in the image and phrases that describe the objects. This is achieved within an open world paradigm, in which the objects in the input image may not have been encountered during the training of the localization mechanism. Moreover, training takes place in a weakly supervised setting, where no bounding boxes are provided. To achieve this, our method combines two pre-trained networks: the CLIP image-to-text matching score and the BLIP image captioning tool. Training takes place on COCO images and their captions and is based on CLIP. Then, during inference, BLIP is used to generate a hypothesis regarding various regions of the current image. Our work generalizes weakly supervised segmentation and phrase grounding and is shown empirically to outperform the state of the art in both domains. It also shows very convincing results in the novel task of weakly-supervised open-world purely visual phrase-grounding presented in our work. For example, on the datasets used for benchmarking phrase-grounding, our method results in a very modest degradation in comparison to methods that employ human captions as an additional input. Our code is available at https://github.com/talshaharabany/what-is-where-by-looking and a live demo can be found at https://replicate.com/talshaharabany/what-is-where-by-looking.
Industrial X-ray analysis is common in aerospace, automotive or nuclear industries where structural integrity of some parts needs to be guaranteed. However, the interpretation of radiographic images is sometimes difficult and may lead to two experts disagree on defect classification. The Automated Defect Recognition (ADR) system presented herein will reduce the analysis time and will also help reducing the subjective interpretation of the defects while increasing the reliability of the human inspector. Our Convolutional Neural Network (CNN) model achieves 94.2\% accuracy (mAP@IoU=50\%), which is considered as similar to expected human performance, when applied to an automotive aluminium castings dataset (GDXray), exceeding current state of the art for this dataset. On an industrial environment, its inference time is less than 400 ms per DICOM image, so it can be installed on production facilities with no impact on delivery time. In addition, an ablation study of the main hyper-parameters to optimise model accuracy from the initial baseline result of 75\% mAP up to 94.2\% mAP, was also conducted.