Geometric feature learning for 3D meshes is central to computer graphics and highly important for numerous vision applications. However, deep learning currently lags in hierarchical modeling of heterogeneous 3D meshes due to the lack of required operations and/or their efficient implementations. In this paper, we propose a series of modular operations for effective geometric deep learning over heterogeneous 3D meshes. These operations include mesh convolutions, (un)pooling and efficient mesh decimation. We provide open source implementation of these operations, collectively termed \textit{Picasso}. The mesh decimation module of Picasso is GPU-accelerated, which can process a batch of meshes on-the-fly for deep learning. Our (un)pooling operations compute features for newly-created neurons across network layers of varying resolution. Our mesh convolutions include facet2vertex, vertex2facet, and facet2facet convolutions that exploit vMF mixture and Barycentric interpolation to incorporate fuzzy modelling. Leveraging the modular operations of Picasso, we contribute a novel hierarchical neural network, PicassoNet-II, to learn highly discriminative features from 3D meshes. PicassoNet-II accepts primitive geometrics and fine textures of mesh facets as input features, while processing full scene meshes. Our network achieves highly competitive performance for shape analysis and scene parsing on a variety of benchmarks. We release Picasso and PicassoNet-II on Github https://github.com/EnyaHermite/Picasso.
Deep Learning (DL) is the most widely used tool in the contemporary field of computer vision. Its ability to accurately solve complex problems is employed in vision research to learn deep neural models for a variety of tasks, including security critical applications. However, it is now known that DL is vulnerable to adversarial attacks that can manipulate its predictions by introducing visually imperceptible perturbations in images and videos. Since the discovery of this phenomenon in 2013~[1], it has attracted significant attention of researchers from multiple sub-fields of machine intelligence. In [2], we reviewed the contributions made by the computer vision community in adversarial attacks on deep learning (and their defenses) until the advent of year 2018. Many of those contributions have inspired new directions in this area, which has matured significantly since witnessing the first generation methods. Hence, as a legacy sequel of [2], this literature review focuses on the advances in this area since 2018. To ensure authenticity, we mainly consider peer-reviewed contributions published in the prestigious sources of computer vision and machine learning research. Besides a comprehensive literature review, the article also provides concise definitions of technical terminologies for non-experts in this domain. Finally, this article discusses challenges and future outlook of this direction based on the literature reviewed herein and [2].
Accurate perception of the surrounding scene is helpful for robots to make reasonable judgments and behaviours. Therefore, developing effective scene representation and recognition methods are of significant importance in robotics. Currently, a large body of research focuses on developing novel auxiliary features and networks to improve indoor scene recognition ability. However, few of them focus on directly constructing object features and relations for indoor scene recognition. In this paper, we analyze the weaknesses of current methods and propose an Object-to-Scene (OTS) method, which extracts object features and learns object relations to recognize indoor scenes. The proposed OTS first extracts object features based on the segmentation network and the proposed object feature aggregation module (OFAM). Afterwards, the object relations are calculated and the scene representation is constructed based on the proposed object attention module (OAM) and global relation aggregation module (GRAM). The final results in this work show that OTS successfully extracts object features and learns object relations from the segmentation network. Moreover, OTS outperforms the state-of-the-art methods by more than 2\% on indoor scene recognition without using any additional streams. Code is publicly available at: https://github.com/FreeformRobotics/OTS.
We propose a self-supervised spatio-temporal matching method coined Motion-Aware Mask Propagation (MAMP) for semi-supervised video object segmentation. During training, MAMP leverages the frame reconstruction task to train the model without the need for annotations. During inference, MAMP extracts high-resolution features from each frame to build a memory bank from the features as well as the predicted masks of selected past frames. MAMP then propagates the masks from the memory bank to subsequent frames according to our motion-aware spatio-temporal matching module, also proposed in this paper. Evaluation on DAVIS-2017 and YouTube-VOS datasets show that MAMP achieves state-of-the-art performance with stronger generalization ability compared to existing self-supervised methods, i.e. 4.9\% higher mean $\mathcal{J}\&\mathcal{F}$ on DAVIS-2017 and 4.85\% higher mean $\mathcal{J}\&\mathcal{F}$ on the unseen categories of YouTube-VOS than the nearest competitor. Moreover, MAMP performs on par with many supervised video object segmentation methods. Our code is available at: \url{https://github.com/bo-miao/MAMP}.
Deep learning is found to be vulnerable to adversarial examples. However, its adversarial susceptibility in image caption generation is under-explored. We study adversarial examples for vision and language models, which typically adopt an encoder-decoder framework consisting of two major components: a Convolutional Neural Network (i.e., CNN) for image feature extraction and a Recurrent Neural Network (RNN) for caption generation. In particular, we investigate attacks on the visual encoder's hidden layer that is fed to the subsequent recurrent network. The existing methods either attack the classification layer of the visual encoder or they back-propagate the gradients from the language model. In contrast, we propose a GAN-based algorithm for crafting adversarial examples for neural image captioning that mimics the internal representation of the CNN such that the resulting deep features of the input image enable a controlled incorrect caption generation through the recurrent network. Our contribution provides new insights for understanding adversarial attacks on vision systems with language component. The proposed method employs two strategies for a comprehensive evaluation. The first examines if a neural image captioning system can be misled to output targeted image captions. The second analyzes the possibility of keywords into the predicted captions. Experiments show that our algorithm can craft effective adversarial images based on the CNN hidden layers to fool captioning framework. Moreover, we discover the proposed attack to be highly transferable. Our work leads to new robustness implications for neural image captioning.
Backpropagation image saliency aims at explaining model predictions by estimating model-centric importance of individual pixels in the input. However, class-insensitivity of the earlier layers in a network only allows saliency computation with low resolution activation maps of the deeper layers, resulting in compromised image saliency. Remedifying this can lead to sanity failures. We propose CAMERAS, a technique to compute high-fidelity backpropagation saliency maps without requiring any external priors and preserving the map sanity. Our method systematically performs multi-scale accumulation and fusion of the activation maps and backpropagated gradients to compute precise saliency maps. From accurate image saliency to articulation of relative importance of input features for different models, and precise discrimination between model perception of visually similar objects, our high-resolution mapping offers multiple novel insights into the black-box deep visual models, which are presented in the paper. We also demonstrate the utility of our saliency maps in adversarial setup by drastically reducing the norm of attack signals by focusing them on the precise regions identified by our maps. Our method also inspires new evaluation metrics and a sanity check for this developing research direction. Code is available here https://github.com/VisMIL/CAMERAS
Deep visual models are susceptible to adversarial perturbations to inputs. Although these signals are carefully crafted, they still appear noise-like patterns to humans. This observation has led to the argument that deep visual representation is misaligned with human perception. We counter-argue by providing evidence of human-meaningful patterns in adversarial perturbations. We first propose an attack that fools a network to confuse a whole category of objects (source class) with a target label. Our attack also limits the unintended fooling by samples from non-sources classes, thereby circumscribing human-defined semantic notions for network fooling. We show that the proposed attack not only leads to the emergence of regular geometric patterns in the perturbations, but also reveals insightful information about the decision boundaries of deep models. Exploring this phenomenon further, we alter the `adversarial' objective of our attack to use it as a tool to `explain' deep visual representation. We show that by careful channeling and projection of the perturbations computed by our method, we can visualize a model's understanding of human-defined semantic notions. Finally, we exploit the explanability properties of our perturbations to perform image generation, inpainting and interactive image manipulation by attacking adversarialy robust `classifiers'.In all, our major contribution is a novel pragmatic adversarial attack that is subsequently transformed into a tool to interpret the visual models. The article also makes secondary contributions in terms of establishing the utility of our attack beyond the adversarial objective with multiple interesting applications.
In recent times, the scope of LIDAR (Light Detection and Ranging) sensor-based technology has spread across numerous fields. It is popularly used to map terrain and navigation information into reliable 3D point cloud data, potentially revolutionizing the autonomous vehicles and assistive robotic industry. A point cloud is a dense compilation of spatial data in 3D coordinates. It plays a vital role in modeling complex real-world scenes since it preserves structural information and avoids perspective distortion, unlike image data, which is the projection of a 3D structure on a 2D plane. In order to leverage the intrinsic capabilities of the LIDAR data, we propose a PointNet-based approach for 3D Multi-Object Tracking (MOT).
Deep Neural Networks lead the state of the art of computer vision tasks. Despite this, Neural Networks are brittle in that small changes in the input can drastically affect their prediction outcome and confidence. Consequently and naturally, research in this area mainly focus on adversarial attacks and defenses. In this paper, we take an alternative stance and introduce the concept of Assistive Signals, which are optimized to improve a model's confidence score regardless if it's under attack or not. We analyse some interesting properties of these assistive perturbations and extend the idea to optimize assistive signals in the 3D space for real-life scenarios simulating different lighting conditions and viewing angles. Experimental evaluations show that the assistive signals generated by our optimization method increase the accuracy and confidence of deep models more than those generated by conventional methods that work in the 2D space. In addition, our Assistive Signals illustrate the intrinsic bias of ML models towards certain patterns in real-life objects. We discuss how we can exploit these insights to re-think, or avoid, some patterns that might contribute to, or degrade, the detectability of objects in the real-world.
3D object grounding aims to locate the most relevant target object in a raw point cloud scene based on a free-form language description. Understanding complex and diverse descriptions, and lifting them directly to a point cloud is a new and challenging topic due to the irregular and sparse nature of point clouds. There are three main challenges in 3D object grounding: to find the main focus in the complex and diverse description; to understand the point cloud scene; and to locate the target object. In this paper, we address all three challenges. Firstly, we propose a language scene graph module to capture the rich structure and long-distance phrase correlations. Secondly, we introduce a multi-level 3D proposal relation graph module to extract the object-object and object-scene co-occurrence relationships, and strengthen the visual features of the initial proposals. Lastly, we develop a description guided 3D visual graph module to encode global contexts of phrases and proposals by a nodes matching strategy. Extensive experiments on challenging benchmark datasets (ScanRefer and Nr3D) show that our algorithm outperforms existing state-of-the-art. Our code is available at https://github.com/PNXD/FFL-3DOG.