On-device machine learning (ML) moves computation from the cloud to personal devices, protecting user privacy and enabling intelligent user experiences. However, fitting models on devices with limited resources presents a major technical challenge: practitioners need to optimize models and balance hardware metrics such as model size, latency, and power. To help practitioners create efficient ML models, we designed and developed Talaria : a model visualization and optimization system. Talaria enables practitioners to compile models to hardware, interactively visualize model statistics, and simulate optimizations to test the impact on inference metrics. Since its internal deployment two years ago, we have evaluated Talaria using three methodologies: (1) a log analysis highlighting its growth of 800+ practitioners submitting 3,600+ models; (2) a usability survey with 26 users assessing the utility of 20 Talaria features; and (3) a qualitative interview with the 7 most active users about their experience using Talaria.

We introduce PointConvFormer, a novel building block for point cloud based deep network architectures. Inspired by generalization theory, PointConvFormer combines ideas from point convolution, where filter weights are only based on relative position, and Transformers which utilize feature-based attention. In PointConvFormer, attention computed from feature difference between points in the neighborhood is used to modify the convolutional weights at each point. Hence, we preserved the invariances from point convolution, whereas attention helps to select relevant points in the neighborhood for convolution. PointConvFormer is suitable for multiple tasks that require details at the point level, such as segmentation and scene flow estimation tasks. We experiment on both tasks with multiple datasets including ScanNet, SemanticKitti, FlyingThings3D and KITTI. Our results show that PointConvFormer offers a better accuracy-speed tradeoff than classic convolutions, regular transformers, and voxelized sparse convolution approaches. Visualizations show that PointConvFormer performs similarly to convolution on flat areas, whereas the neighborhood selection effect is stronger on object boundaries, showing that it has got the best of both worlds. The code will be available.

Implicit neural fields, typically encoded by a multilayer perceptron (MLP) that maps from coordinates (e.g., xyz) to signals (e.g., signed distances), have shown remarkable promise as a high-fidelity and compact representation. However, the lack of a regular and explicit grid structure also makes it challenging to apply generative modeling directly on implicit neural fields in order to synthesize new data. To this end, we propose HyperDiffusion, a novel approach for unconditional generative modeling of implicit neural fields. HyperDiffusion operates directly on MLP weights and generates new neural implicit fields encoded by synthesized MLP parameters. Specifically, a collection of MLPs is first optimized to faithfully represent individual data samples. Subsequently, a diffusion process is trained in this MLP weight space to model the underlying distribution of neural implicit fields. HyperDiffusion enables diffusion modeling over a implicit, compact, and yet high-fidelity representation of complex signals across 3D shapes and 4D mesh animations within one single unified framework.

As neural networks grow in size and complexity, inference speeds decline. To combat this, one of the most effective compression techniques -- channel pruning -- removes channels from weights. However, for multi-branch segments of a model, channel removal can introduce inference-time memory copies. In turn, these copies increase inference latency -- so much so that the pruned model can be slower than the unpruned model. As a workaround, pruners conventionally constrain certain channels to be pruned together. This fully eliminates memory copies but, as we show, significantly impairs accuracy. We now have a dilemma: Remove constraints but increase latency, or add constraints and impair accuracy. In response, our insight is to reorder channels at export time, (1) reducing latency by reducing memory copies and (2) improving accuracy by removing constraints. Using this insight, we design a generic algorithm UPSCALE to prune models with any pruning pattern. By removing constraints from existing pruners, we improve ImageNet accuracy for post-training pruned models by 2.1 points on average -- benefiting DenseNet (+16.9), EfficientNetV2 (+7.9), and ResNet (+6.2). Furthermore, by reordering channels, UPSCALE improves inference speeds by up to 2x over a baseline export.

Reconstructing an accurate 3D object model from a few image observations remains a challenging problem in computer vision. State-of-the-art approaches typically assume accurate camera poses as input, which could be difficult to obtain in realistic settings. In this paper, we present FvOR, a learning-based object reconstruction method that predicts accurate 3D models given a few images with noisy input poses. The core of our approach is a fast and robust multi-view reconstruction algorithm to jointly refine 3D geometry and camera pose estimation using learnable neural network modules. We provide a thorough benchmark of state-of-the-art approaches for this problem on ShapeNet. Our approach achieves best-in-class results. It is also two orders of magnitude faster than the recent optimization-based approach IDR [67].

Deep learning has made significant impacts on multiview stereo systems. State-of-the-art approaches typically involve building a cost volume, followed by multiple 3D convolution operations to recover the input image's pixel-wise depth. While such end-to-end learning of plane-sweeping stereo advances public benchmarks' accuracy, they are typically very slow to compute. We present MVS2D, a highly efficient multi-view stereo algorithm that seamlessly integrates multi-view constraints into single-view net-works via an attention mechanism. Since MVS2D only builds on 2D convolutions, it is at least$2\times faster$than all the notable counterparts. Moreover, our algorithm produces precise depth estimations and 3D reconstructions, achieving state-of-the-art results on challenging benchmarks ScanNet, SUN3D, RGBD, and the classical DTU dataset. our algorithm also outperforms all other algorithms in the setting of inexact camera poses. Our code is released at https://github.com/zhenpeiyang/MVS2D

We study the problem of novel view synthesis from sparse source observations of a scene comprised of 3D objects. We propose a simple yet effective approach that is neither continuous nor implicit, challenging recent trends on view synthesis. Our approach explicitly encodes observations into a volumetric representation that enables amortized rendering. We demonstrate that although continuous radiance field representations have gained a lot of attention due to their expressive power, our simple approach obtains comparable or even better novel view reconstruction quality comparing with state-of-the-art baselines [49] while increasing rendering speed by over 400x. Our model is trained in a category-agnostic manner and does not require scene-specific optimization. Therefore, it is able to generalize novel view synthesis to object categories not seen during training. In addition, we show that with our simple formulation, we can use view synthesis as a self-supervision signal for efficient learning of 3D geometry without explicit 3D supervision.

Many accessibility features available on mobile platforms require applications (apps) to provide complete and accurate metadata describing user interface (UI) components. Unfortunately, many apps do not provide sufficient metadata for accessibility features to work as expected. In this paper, we explore inferring accessibility metadata for mobile apps from their pixels, as the visual interfaces often best reflect an app’s full functionality. We trained a robust, fast, memory-efficient, on-device model to detect UI elements using a dataset of 77,637 screens (from 4,068 iPhone apps) that we collected and annotated. To further improve UI detections and add semantic information, we introduced heuristics (e.g., UI grouping and ordering) and additional models (e.g., recognize UI content, state, interactivity). We built Screen Recognition to generate accessibility metadata to augment iOS VoiceOver. In a study with 9 screen reader users, we validated that our approach improves the accessibility of existing mobile apps, enabling even previously inaccessible apps to be used.

3D reconstruction of large scenes is a challenging problem due to the high-complexity nature of the solution space, in particular for generative neural networks. In contrast to traditional generative learned models which encode the full generative process into a neural network and can struggle with maintaining local details at the scene level, we introduce a new method that directly leverages scene geometry from the training database. First, we learn to synthesize an initial estimate for a 3D scene, constructed by retrieving a top-k set of volumetric chunks from the scene database. These candidates are then refined to a final scene generation with an attention-based refinement that can effectively select the most consistent set of geometry from the candidates and combine them together to create an output scene, facilitating transfer of coherent structures and local detail from train scene geometry. We demonstrate our neural scene reconstruction with a database for the tasks of 3D super resolution and surface reconstruction from sparse point clouds, showing that our approach enables generation of more coherent, accurate 3D scenes, improving on average by over 8% in IoU over state-of-the-art scene reconstruction.

Reading and writing are the foundations of English learning as well as an important method of instruction. With the advancement of network technology and the onset of the information age, an increasing number of students have lost interest in traditional English reading and writing instruction in the classroom. Flipped classrooms have emerged as a result of this situation and have become the focus of research in one fell swoop. As a result, flipped classroom research at home and abroad has primarily focused on the theory and practical application of flipped classrooms, and flipped classroom application practice is primarily based on the overall classroom, with few separate discussions on the effects of flipped classroom students’ self‐learning. As a result, we developed a recurrent neural network‐based intelligent assisted learning algorithm for English flipped classrooms. There are two main characteristics of the model. First, it is a gated recurrent unit based on a variant structure of the recurrent neural network. The double‐gating mechanism fully considers the context and selects memory through weight assignment, and on this basis, it integrates the novel LeakyReLU function to improve the model’s training convergence efficiency. Second, by overcoming time‐consuming problems in the medium, the adoption of the connection sequence classification algorithm eliminates the need for prior alignment of speech and text data, resulting in a direct boost in model training speed. The experimental results show that in the English flipped classroom’s intelligent learning mode, students explore and discover knowledge independently, their enthusiasm and interest in learning are greatly increased, and the flipped classroom’s teaching effect is greatly improved.

We propose a framework for learning neural scene representations directly from images, without 3D supervision. Our key insight is that 3D structure can be imposed by ensuring that the learned representation transforms like a real 3D scene. Specifically, we introduce a loss which enforces equivariance of the scene representation with respect to 3D transformations. Our formulation allows us to infer and render scenes in real time while achieving comparable results to models requiring minutes for inference. In addition, we introduce two challenging new datasets for scene representation and neural rendering, including scenes with complex lighting and backgrounds. Through experiments, we show that our model achieves compelling results on these datasets as well as on standard ShapeNet benchmarks.

We propose an algorithm to predict room layout from a single image that generalizes across panoramas and perspective images, cuboid layouts and more general layouts (e.g. "L"-shape room). Our method operates directly on the panoramic image, rather than decomposing into perspective images as do recent works. Our network architecture is similar to that of RoomNet [15], but we show improvements due to aligning the image based on vanishing points, predicting multiple layout elements (corners, boundaries, size and translation), and fitting a constrained Manhattan layout to the resulting predictions. Our method compares well in speed and accuracy to other existing work on panoramas, achieves among the best accuracy for perspective images, and can handle both cuboid-shaped and more general Manhattan layouts.

Given a single photo of a room and a large database of furniture CAD models, our goal is to reconstruct a scene that is as similar as possible to the scene depicted in the photograph, and composed of objects drawn from the database. We present a completely automatic system to address this IM2CAD problem that produces high quality results on challenging imagery from interior home design and remodeling websites. Our approach iteratively optimizes the placement and scale of objects in the room to best match scene renderings to the input photo, using image comparison metrics trained via deep convolutional neural nets. By operating jointly on the full scene at once, we account for inter-object occlusions. We also show the applicability of our method in standard scene understanding benchmarks where we obtain significant improvement.

Background: Code inspection has been accepted as an effective method to detect and remove defects and code reading is a critical step in code inspection. However, there are very limited empirical studies on the content and appropriate forms of the suitable software artifacts as the supporting materials, hence inspectors may not be well-supported with necessary knowledge to carry out code reading. Objective: This research aims to investigate the impact of different common supporting materials (i.e., comments vs. design documents) on code reading. Method: A relatively large-scale controlled experiment with 135 senior students was designed and executed to compare the impacts of different supporting materials on code reading. The subjects were randomly separated into three groups with different treatments, i.e, the comments, the design documents and the comments+design documents, respectively. Two metrics regarding the code reading performance (i.e., Effectiveness and Defect Detection Rate) were used to compare the different impacts derived from the two different types of supporting materials. Qualitative feedbacks were also collected using questionnaires for the final analysis. Results: The results indicate that students performed better when being provided with comments than comments+design documents. Also, the removal of design documents shows little impact on inspection effectiveness and may lead to an increase in defect detection rate. Conclusion: Comments may provide more help and value than design documents as supporting material in small to median sized code reading.

Context: Capture-recapture method has long been adopted in software engineering as a relatively objective way for defect estimation. While many relevant studies have been carried out to evaluate various capture-recapture models and estimators, there still lacks common understanding on the adoption status of the method in software engineering. It is necessary to systematically collect empirical evidence of Capture-recapture adoption hence form necessary understanding on the method.

We address the problem of geo-registering ground-based multi-view stereo models by ground-to-aerial image matching. The main contribution is a fully automated geo-registration pipeline with a novel viewpoint-dependent matching method that handles ground to aerial viewpoint variation. We conduct large-scale experiments which consist of many popular outdoor landmarks in Rome. The proposed approach demonstrates a high success rate for the task, and dramatically outperforms state-of-the-art techniques, yielding geo-registration at pixel-level accuracy.

This paper leverages occluding contours (aka "internal silhouettes") to improve the performance of multi-view stereo methods. The contributions are 1) a new technique to identify free-space regions arising from occluding contours, and 2) a new approach for incorporating the resulting free-space constraints into Poisson surface reconstruction. The proposed approach outperforms state of the art MVS techniques for challenging Internet datasets, yielding dramatic quality improvements both around object contours and in surface detail.

We present the first large scale system for capturing and rendering relight able scene reconstructions from massive unstructured photo collections taken under different illumination conditions and viewpoints. We combine photos taken from many sources, Flickr-Based ground-level imagery, oblique aerial views, and street view, to recover models that are significantly more complete and detailed than previously demonstrated. We demonstrate the ability to match both the viewpoint and illumination of arbitrary input photos, enabling a Visual Turing Test in which photo and rendering are viewed side-by-side and the observer has to guess which is which. While we cannot yet fool human perception, the gap is closing.

We propose a novel framework for reconstructing homogenous, transparent, refractive height-fields from a single viewpoint. The height-field is imaged against a known planar background, or sequence of backgrounds. Unlike existing approaches that do a point-by-point reconstruction - which is known to have intractable ambiguities - our method estimates and optimizes for the entire height-field at the same time. The formulation supports shape recovery from measured distortions (deflections) or directly from the images themselves, including from a single image. We report results for a variety of refractive height-fields showing significant improvement over prior art.

Effective reduction of noise is generally difficult because of the possible tight coupling of noise with high-frequency image structure. The problem is worse under low-light conditions. In this paper, we propose slightly optically defocusing the image in order to loosen this noise-image structure coupling. This allows us to more effectively reduce noise and subsequently restore the small defocus. We analytically show how this is possible, and demonstrate our technique on a number of examples that include low-light images.

This paper introduces a new tone mapping operator that performs local linear adjustments on small overlapping windows over the entire input image. While each window applies a local linear adjustment that preserves the monotonicity of the radiance values, the problem is implicitly cast as one of global optimization that satisfies the local constraints defined on each of the overlapping windows. Local constraints take the form of a guidance map that can be used to effectively suppress local high contrast while preserving details. Using this method, image structures can be preserved even in challenging high dynamic range (HDR) images that contain either abrupt radiance change, or relatively smooth but salient transitions. Another benefit of our formulation is that it can be used to synthesize HDR images from low dynamic range (LDR) images.

We present a new algorithm for removing motion blur from a single image. Our method computes a deblurred image using a unified probabilistic model of both blur kernel estimation and unblurred image restoration. We present an analysis of the causes of common artifacts found in current deblurring methods, and then introduce several novel terms within this probabilistic model that are inspired by our analysis. These terms include a model of the spatial randomness of noise in the blurred image, as well a new local smoothness prior that reduces ringing artifacts by constraining contrast in the unblurred image wherever the blurred image exhibits low contrast. Finally, we describe an effficient optimization scheme that alternates between blur kernel estimation and unblurred image restoration until convergence. As a result of these steps, we are able to produce high quality deblurred results in low computation time. We are even able to produce results of comparable quality to techniques that require additional input images beyond a single blurry photograph, and to methods that require additional hardware.

We propose a simple but effective upsampling method for automatically enhancing the image/video resolution, while preserving the essential structural information. The main advantage of our method lies in a feedback-control framework which faithfully recovers the high-resolution image information from the input data,withoutimposing additional local structure constraints learned from other examples. This makes our method independent of the quality and number of the selected examples, which are issues typical of learning-based algorithms, while producing high-quality results without observable unsightly artifacts. Another advantage is that our method naturally extends to video upsampling, where the temporal coherence is maintained automatically. Finally, our method runs very fast. We demonstrate the effectiveness of our algorithm by experimenting with different image/video data.

Most previous motion deblurring methods restore the degraded image assuming a shift-invariant linear blur filter. These methods are not applicable if the blur is caused by spatially variant motions. In this paper, we model the physical properties of a 2-D rigid body movement and propose a practical framework to deblur rotational motions from a single image. Our main observation is that the transparency cue of a blurred object, which represents the motion blur formation from an imaging perspective, provides sufficient information in determining the object movements. Comparatively, single image motion deblurring using pixel color/gradient information has large uncertainties in motion representation and computation. Our results are produced by minimizing a new energy function combining rotation, possible translations, and the transparency map using an iterative optimizing process. The effectiveness of our method is demonstrated using challenging image examples.