This paper summarizes the design, experiments and results of our solution to the Road Damage Detection and Classification Challenge held as part of the 2018 IEEE International Conference On Big Data Cup. Automatic detection and classification of damage in roads is an essential problem for multiple applications like maintenance and autonomous driving. We demonstrate that convolutional neural net based instance detection and classfication approaches can be used to solve this problem. In particular we show that Mask-RCNN, one of the state-of-the-art algorithms for object detection, localization and instance segmentation of natural images, can be used to perform this task in a fast manner with effective results. We achieve a mean F1 score of 0.528 at an IoU of 50% on the task of detection and classification of different types of damages in real-world road images acquired using a smartphone camera and our average inference time for each image is 0.105 seconds on an NVIDIA GeForce 1080Ti graphic card. The code and saved models for our approach can be found here : https://github.com/sshkhr/BigDataCup18 Submission
An instance with a bad mask might make a composite image that uses it look fake. This encourages us to learn segmentation by generating realistic composite images. To achieve this, we propose a novel framework that exploits a new proposed prior called the independence prior based on Generative Adversarial Networks (GANs). The generator produces an image with multiple category-specific instance providers, a layout module and a composition module. Firstly, each provider independently outputs a category-specific instance image with a soft mask. Then the provided instances' poses are corrected by the layout module. Lastly, the composition module combines these instances into a final image. Training with adversarial loss and penalty for mask area, each provider learns a mask that is as small as possible but enough to cover a complete category-specific instance. Weakly supervised semantic segmentation methods widely use grouping cues modeling the association between image parts, which are either artificially designed or learned with costly segmentation labels or only modeled on local pairs. Unlike them, our method automatically models the dependence between any parts and learns instance segmentation. We apply our framework in two cases: (1) Foreground segmentation on category-specific images with box-level annotation. (2) Unsupervised learning of instance appearances and masks with only one image of homogeneous object cluster (HOC). We get appealing results in both tasks, which shows the independence prior is useful for instance segmentation and it is possible to unsupervisedly learn instance masks with only one image.
Tree-like structures, such as blood vessels, often express complexity at very fine scales, requiring high-resolution grids to adequately describe their shape. Such sparse morphology can alternately be represented by locations of centreline points, but learning from this type of data with deep learning is challenging due to it being unordered, and permutation invariant. In this work, we propose a deep neural network that directly consumes unordered points along the centreline of a branching structure, to identify the topology of the represented structure in a single-shot. Key to our approach is the use of a novel multi-task loss function, enabling instance segmentation of arbitrarily complex branching structures. We train the network solely using synthetically generated data, utilizing domain randomization to facilitate the transfer to real 2D and 3D data. Results show that our network can reliably extract meaningful information about branch locations, bifurcations and endpoints, and sets a new benchmark for semantic instance segmentation in branching structures.
Varying weather conditions, including rainfall and snowfall, are generally regarded as a challenge for computer vision algorithms. One proposed solution to the challenges induced by rain and snowfall is to artificially remove the rain from images or video using rain removal algorithms. It is the promise of these algorithms that the rain-removed image frames will improve the performance of subsequent segmentation and tracking algorithms. However, rain removal algorithms are typically evaluated on their ability to remove synthetic rain on a small subset of images. Currently, their behavior is unknown on real-world videos when integrated with a typical computer vision pipeline. In this paper, we review the existing rain removal algorithms and propose a new dataset that consists of 22 traffic surveillance sequences under a broad variety of weather conditions that all include either rain or snowfall. We propose a new evaluation protocol that evaluates the rain removal algorithms on their ability to improve the performance of subsequent segmentation, instance segmentation, and feature tracking algorithms under rain and snow. If successful, the de-rained frames of a rain removal algorithm should improve segmentation performance and increase the number of accurately tracked features. The results show that a recent single-frame-based rain removal algorithm increases the segmentation performance by 19.7% on our proposed dataset, but it eventually decreases the feature tracking performance and showed mixed results with recent instance segmentation methods. However, the best video-based rain removal algorithm improves the feature tracking accuracy by 7.72%.
In this work, we propose a mask propagation network to treat the video segmentation problem as a concept of the guided instance segmentation. Similar to most MaskTrack based video segmentation methods, our method takes the mask probability map of previous frame and the appearance of current frame as inputs, and predicts the mask probability map for the current frame. Specifically, we adopt the Xception backbone based DeepLab v3+ model as the probability map predictor in our prediction pipeline. Besides, instead of the full image and the original mask probability, our network takes the region of interest of the instance, and the new mask probability which warped by the optical flow between the previous and current frames as the inputs. We also ensemble the modified One-Shot Video Segmentation Network to make the final predictions in order to retrieve and segment the missing instance.
The quantitative analysis of cellular membranes helps understanding developmental processes at the cellular level. Particularly 3D microscopic image data offers valuable insights into cell dynamics, but error-free automatic segmentation remains challenging due to the huge amount of data generated and strong variations in image intensities. In this paper, we propose a new 3D segmentation approach which combines the discriminative power of convolutional neural networks (CNNs) for preprocessing and investigates the performance of three watershed-based postprocessing strategies (WS), which are well suited to segment object shapes, even when supplied with vague seed and boundary constraints. To leverage the full potential of the watershed algorithm, the multi-instance segmentation problem is initially interpreted as three-class semantic segmentation problem, which in turn is well-suited for the application of CNNs. Using manually annotated 3D confocal microscopy images of Arabidopsis thaliana, we show the superior performance of the proposed method compared to the state of the art.
In multi-task learning, multiple tasks are solved jointly, sharing inductive bias between them. Multi-task learning is inherently a multi-objective problem because different tasks may conflict, necessitating a trade-off. A common compromise is to optimize a proxy objective that minimizes a weighted linear combination of per-task losses. However, this workaround is only valid when the tasks do not compete, which is rarely the case. In this paper, we explicitly cast multi-task learning as multi-objective optimization, with the overall objective of finding a Pareto optimal solution. To this end, we use algorithms developed in the gradient-based multi-objective optimization literature. These algorithms are not directly applicable to large-scale learning problems since they scale poorly with the dimensionality of the gradients and the number of tasks. We therefore propose an upper bound for the multi-objective loss and show that it can be optimized efficiently. We further prove that optimizing this upper bound yields a Pareto optimal solution under realistic assumptions. We apply our method to a variety of multi-task deep learning problems including digit classification, scene understanding (joint semantic segmentation, instance segmentation, and depth estimation), and multi-label classification. Our method produces higher-performing models than recent multi-task learning formulations or per-task training.
In the recent years, public use of artistic effects for editing and beautifying images has encouraged researchers to look for new approaches to this task. Most of the existing methods apply artistic effects to the whole image. Exploitation of neural network vision technologies like object detection and semantic segmentation could be a new viewpoint in this area. In this paper, we utilize an instance segmentation neural network to obtain a class mask for separately filtering the background and foreground of an image. We implement a top prior-mask selection to let us select an object class for filtering purpose. Different artistic effects are used in the filtering process to meet the requirements of a vast variety of users. Also, our method is flexible enough to allow the addition of new filters. We use pre-trained Mask R-CNN instance segmentation on the COCO dataset as the segmentation network. Experimental results on the use of different filters are performed. System's output results show that this novel approach can create satisfying artistic images with fast operation and simple interface.
We present an auxiliary task to Mask R-CNN, an instance segmentation network, which leads to faster training of the mask head. Our addition to Mask R-CNN is a new prediction head, the Edge Agreement Head, which is inspired by the way human annotators perform instance segmentation. Human annotators copy the contour of an object instance and only indirectly the occupied instance area. Hence, the edges of instance masks are particularly useful as they characterize the instance well. The Edge Agreement Head therefore encourages predicted masks to have similar image gradients to the ground-truth mask using edge detection filters. We provide a detailed survey of loss combinations and show improvements on the MS COCO Mask metrics compared to using no additional loss. Our approach marginally increases the model size and adds no additional trainable model variables. While the computational costs are increased slightly, the increment is negligible considering the high computational cost of the Mask R-CNN architecture. As the additional network head is only relevant during training, inference speed remains unchanged compared to Mask R-CNN. In a default Mask R-CNN setup, we achieve a training speed-up and a relative overall improvement of 8.1% on the MS COCO metrics compared to the baseline.
The ability to segment unknown objects in depth images has potential to enhance robot skills in grasping and object tracking. Recent computer vision research has demonstrated that Mask R-CNN can be trained to segment specific categories of objects in RGB images when massive hand-labeled datasets are available. As generating these datasets is time consuming, we instead train with synthetic depth images. Many robots now use depth sensors, and recent results suggest training on synthetic depth data can transfer successfully to the real world. We present a method for automated dataset generation and rapidly generate a synthetic training dataset of 50,000 depth images and 320,000 object masks using simulated heaps of 3D CAD models. We train a variant of Mask R-CNN with domain randomization on the generated dataset to perform category-agnostic instance segmentation without any hand-labeled data and we evaluate the trained network, which we refer to as Synthetic Depth (SD) Mask R-CNN, on a set of real, high-resolution depth images of challenging, densely-cluttered bins containing objects with highly-varied geometry. SD Mask R-CNN outperforms point cloud clustering baselines by an absolute 15% in Average Precision and 20% in Average Recall on COCO benchmarks, and achieves performance levels similar to a Mask R-CNN trained on a massive, hand-labeled RGB dataset and fine-tuned on real images from the experimental setup. We deploy the model in an instance-specific grasping pipeline to demonstrate its usefulness in a robotics application. Code, the synthetic training dataset, and supplementary material are available at https://bit.ly/2letCuE.
Adversarial noises are useful tools to probe the weakness of deep learning based computer vision algorithms. In this paper, we describe a robust adversarial perturbation (R-AP) method to attack deep proposal-based object detectors and instance segmentation algorithms. Our method focuses on attacking the common component in these algorithms, namely Region Proposal Network (RPN), to universally degrade their performance in a black-box fashion. To do so, we design a loss function that combines a label loss and a novel shape loss, and optimize it with respect to image using a gradient based iterative algorithm. Evaluations are performed on the MS COCO 2014 dataset for the adversarial attacking of 6 state-of-the-art object detectors and 2 instance segmentation algorithms. Experimental results demonstrate the efficacy of the proposed method.
The task of generating natural images from 3D scenes has been a long standing goal in computer graphics. On the other hand, recent developments in deep neural networks allow for trainable models that can produce natural-looking images with little or no knowledge about the scene structure. While the generated images often consist of realistic looking local patterns, the overall structure of the generated images is often inconsistent. In this work we propose a trainable, geometry-aware image generation method that leverages various types of scene information, including geometry and segmentation, to create realistic looking natural images that match the desired scene structure. Our geometrically-consistent image synthesis method is a deep neural network, called Geometry to Image Synthesis (GIS) framework, which retains the advantages of a trainable method, e.g., differentiability and adaptiveness, but, at the same time, makes a step towards the generalizability, control and quality output of modern graphics rendering engines. We utilize the GIS framework to insert vehicles in outdoor driving scenes, as well as to generate novel views of objects from the Linemod dataset. We qualitatively show that our network is able to generalize beyond the training set to novel scene geometries, object shapes and segmentations. Furthermore, we quantitatively show that the GIS framework can be used to synthesize large amounts of training data which proves beneficial for training instance segmentation models.
In this paper, we investigate the use of an unsupervised label clustering technique and demonstrate that it enables substantial improvements in visual relationship prediction accuracy on the Person in Context (PIC) dataset. We propose to group object labels with similar patterns of relationship distribution in the dataset into fewer categories. Label clustering not only mitigates both the large classification space and class imbalance issues, but also potentially increases data samples for each clustered category. We further propose to incorporate depth information as an additional feature into the instance segmentation model. The additional depth prediction path supplements the relationship prediction model in a way that bounding boxes or segmentation masks are unable to deliver. We have rigorously evaluated the proposed techniques and performed various ablation analysis to validate the benefits of them.
Performing data augmentation for learning deep neural networks is known to be important for training visual recognition systems. By artificially increasing the number of training examples, it helps reducing overfitting and improves generalization. While simple image transformations can already improve predictive performance in most vision tasks, larger gains can be obtained by leveraging task-specific prior knowledge. In this work, we consider object detection, semantic and instance segmentation and augment the training images by blending objects in existing scenes, using instance segmentation annotations. We observe that randomly pasting objects on images hurts the performance, unless the object is placed in the right context. To resolve this issue, we propose an explicit context model by using a convolutional neural network, which predicts whether an image region is suitable for placing a given object or not. In our experiments, we show that our approach is able to improve object detection, semantic and instance segmentation on the PASCAL VOC12 and COCO datasets, with significant gains in a limited annotation scenario, i.e. when only one category is annotated. We also show that the method is not limited to datasets that come with expensive pixel-wise instance annotations and can be used when only bounding boxes are available, by employing weakly-supervised learning for instance masks approximation.
We present a single network method for panoptic segmentation. This method combines the predictions from a jointly trained semantic and instance segmentation network using heuristics. Joint training is the first step towards an end-to-end panoptic segmentation network and is faster and more memory efficient than training and predicting with two networks, as done in previous work. The architecture consists of a ResNet-50 feature extractor shared by the semantic segmentation and instance segmentation branch. For instance segmentation, a Mask R-CNN type of architecture is used, while the semantic segmentation branch is augmented with a Pyramid Pooling Module. Results for this method are submitted to the COCO and Mapillary Joint Recognition Challenge 2018. Our approach achieves a PQ score of 17.6 on the Mapillary Vistas validation set and 27.2 on the COCO test-dev set.
We present a weakly supervised model that jointly performs both semantic- and instance-segmentation -- a particularly relevant problem given the substantial cost of obtaining pixel-perfect annotation for these tasks. In contrast to many popular instance segmentation approaches based on object detectors, our method does not predict any overlapping instances. Moreover, we are able to segment both "thing" and "stuff" classes, and thus explain all the pixels in the image. "Thing" classes are weakly-supervised with bounding boxes, and "stuff" with image-level tags. We obtain state-of-the-art results on Pascal VOC, for both full and weak supervision (which achieves about 95% of fully-supervised performance). Furthermore, we present the first weakly-supervised results on Cityscapes for both semantic- and instance-segmentation. Finally, we use our weakly supervised framework to analyse the relationship between annotation quality and predictive performance, which is of interest to dataset creators.
Quality control is a fundamental component of many manufacturing processes, especially those involving casting or welding. However, manual quality control procedures are often time-consuming and error-prone. In order to meet the growing demand for high-quality products, the use of intelligent visual inspection systems is becoming essential in production lines. Recently, Convolutional Neural Networks (CNNs) have shown outstanding performance in both image classification and localization tasks. In this article, a system is proposed for the identification of casting defects in X-ray images, based on the Mask Region-based CNN architecture. The proposed defect detection system simultaneously performs defect detection and segmentation on input images, making it suitable for a range of defect detection tasks. It is shown that training the network to simultaneously perform defect detection and defect instance segmentation, results in a higher defect detection accuracy than training on defect detection alone. Transfer learning is leveraged to reduce the training data demands and increase the prediction accuracy of the trained model. More specifically, the model is first trained with two large openly-available image datasets before finetuning on a relatively small metal casting X-ray dataset. The accuracy of the trained model exceeds state-of-the art performance on the GRIMA database of X-ray images (GDXray) Castings dataset and is fast enough to be used in a production setting. The system also performs well on the GDXray Welds dataset. A number of in-depth studies are conducted to explore how transfer learning, multi-task learning, and multi-class learning influence the performance of the trained system.
Despite the recent advancements in deploying neural networks for image classification, it has been found that adversarial examples are able to fool these models leading them to misclassify the images. Since these models are now being widely deployed, we provide an insight on the threat of these adversarial examples by evaluating their characteristics and transferability to more complex models that utilize Image Classification as a subtask. We demonstrate the ineffectiveness of adversarial examples when applied to Instance Segmentation & Object Detection models. We show that this ineffectiveness arises from the inability of adversarial examples to withstand transformations such as scaling or a change in lighting conditions. Moreover, we show that there exists a small threshold below which the adversarial property is retained while applying these input transformations. Additionally, these attacks demonstrate weak cross-network transferability across neural network architectures, e.g. VGG16 and ResNet50, however, the attack may fool both the networks if passed sequentially through networks during its formation. The lack of scalability and transferability challenges the question of how adversarial images would be effective in the real world.
Automated segmentation of individual leaves of a plant in an image is a prerequisite to measure more complex phenotypic traits in high-throughput phenotyping. Applying state-of-the-art machine learning approaches to tackle leaf instance segmentation requires a large amount of manually annotated training data. Currently, the benchmark datasets for leaf segmentation contain only a few hundred labeled training images. In this paper, we propose a framework for leaf instance segmentation by augmenting real plant datasets with generated synthetic images of plants inspired by domain randomisation. We train a state-of-the-art deep learning segmentation architecture (Mask-RCNN) with a combination of real and synthetic images of Arabidopsis plants. Our proposed approach achieves 90% leaf segmentation score on the A1 test set outperforming the-state-of-the-art approaches for the CVPPP Leaf Segmentation Challenge (LSC). Our approach also achieves 81% mean performance over all five test datasets.
Object detection and instance segmentation are dominated by region-based methods such as Mask RCNN. However, there is a growing interest in reducing these problems to pixel labeling tasks, as the latter could be more efficient, could be integrated seamlessly in image-to-image network architectures as used in many other tasks, and could be more accurate for objects that are not well approximated by bounding boxes. In this paper we show theoretically and empirically that constructing dense pixel embeddings that can separate object instances cannot be easily achieved using convolutional operators. At the same time, we show that simple modifications, which we call semi-convolutional, have a much better chance of succeeding at this task. We use the latter to show a connection to Hough voting as well as to a variant of the bilateral kernel that is spatially steered by a convolutional network. We demonstrate that these operators can also be used to improve approaches such as Mask RCNN, demonstrating better segmentation of complex biological shapes and PASCAL VOC categories than achievable by Mask RCNN alone.