Image segmentation, one of the most critical vision tasks, has been studied for many years. Most of the early algorithms are unsupervised methods, which use hand-crafted features to divide the image into many regions. Recently, owing to the great success of deep learning technology, CNNs based methods show superior performance in image segmentation. However, these methods rely on a large number of human annotations, which are expensive to collect. In this paper, we propose a deep unsupervised method for image segmentation, which contains the following two stages. First, a Superpixelwise Autoencoder (SuperAE) is designed to learn the deep embedding and reconstruct a smoothed image, then the smoothed image is passed to generate superpixels. Second, we present a novel clustering algorithm called Deep Superpixel Cut (DSC), which measures the deep similarity between superpixels and formulates image segmentation as a soft partitioning problem. Via backpropagation, DSC adaptively partitions the superpixels into perceptual regions. Experimental results on the BSDS500 dataset demonstrate the effectiveness of the proposed method.
Superpixels are a useful representation to reduce the complexity of image data. However, to combine superpixels with convolutional neural networks (CNNs) in an end-to-end fashion, one requires extra models to generate superpixels and special operations such as graph convolution. In this paper, we propose a way to implicitly integrate a superpixel scheme into CNNs, which makes it easy to use superpixels with CNNs in an end-to-end fashion. Our proposed method hierarchically groups pixels at downsampling layers and generates superpixels. Our method can be plugged into many existing architectures without a change in their feed-forward path because our method does not use superpixels in the feed-forward path but use them to recover the lost resolution instead of bilinear upsampling. As a result, our method preserves detailed information such as object boundaries in the form of superpixels even when the model contains downsampling layers. We evaluate our method on several tasks such as semantic segmentation, superpixel segmentation, and monocular depth estimation, and confirm that it speeds up modern architectures and/or improves their prediction accuracy in these tasks.
We propose a new fast fully unsupervised method to discover semantic patterns. Our algorithm is able to hierarchically find visual categories and produce a segmentation mask where previous methods fail. Through the modeling of what is a visual pattern in an image, we introduce the notion of "semantic levels" and devise a conceptual framework along with measures and a dedicated benchmark dataset for future comparisons. Our algorithm is composed by two phases. A filtering phase, which selects semantical hotsposts by means of an accumulator space, then a clustering phase which propagates the semantic properties of the hotspots on a superpixels basis. We provide both qualitative and quantitative experimental validation, achieving optimal results in terms of robustness to noise and semantic consistency. We also made code and dataset publicly available.
The performance of modern algorithms on certain computer vision tasks such as object recognition is now close to that of humans. This success was achieved at the price of complicated architectures depending on millions of parameters and it has become quite challenging to understand how particular predictions are made. Interpretability methods propose to give us this understanding. In this paper, we study LIME, perhaps one of the most popular. On the theoretical side, we show that when the number of generated examples is large, LIME explanations are concentrated around a limit explanation for which we give an explicit expression. We further this study for elementary shape detectors and linear models. As a consequence of this analysis, we uncover a connection between LIME and integrated gradients, another explanation method. More precisely, the LIME explanations are similar to the sum of integrated gradients over the superpixels used in the preprocessing step of LIME.
Several approaches have been proposed to solve the spectral unmixing problem in hyperspectral image analysis. Among them the use of sparse regression techniques aims to characterize the abundances in pixels based on a large library of spectral signatures known a priori. Recently, the integration of image spatial-contextual information significantly enhanced the performance of sparse unmixing. In this work, we propose a computationally efficient multiscale representation method for hyperspectral data adapted to the unmixing problem. The proposed method is based on a hierarchical extension of the SLIC oversegmentation algorithm constructed using a robust homogeneity testing. The image is subdivided into a set of spectrally homogeneous regions formed by pixels with similar characteristics (superpixels). This representation is then used to provide prior spatial regularity information for the abundances of materials present in the scene, improving the conditioning of the unmixing problem. Simulation results illustrate that the method is capable of estimating abundances with high quality and low computational cost, especially in noisy scenarios.
Superpixel algorithms grouping pixels with similar color and other low-level properties are increasingly used for pre-processing in image segmentation. In recent years, a focus has been placed on developing geometric superpixel methods that facilitate the extraction and analysis of geometric image features. Diagram-based superpixel methods are important among the geometric methods as they generate compact and sparsely representable superpixels. Introducing generalized balanced power diagrams to the field of superpixels, we propose a diagram method called Power-SLIC. Power-SLIC is the first geometric superpixel method to generate piecewise quadratic boundaries. Its speed, competitive with fast state-of-the-art methods, is unprecedented for diagram approaches. Extensive computational experiments show that Power-SLIC outperforms existing diagram approaches in boundary recall, under segmentation error, achievable segmentation accuracy, and compression quality. Moreover, Power-SLIC is robust to Gaussian noise.
Hyperspectral image classification (HIC) is an important but challenging task, and a problem that limits the algorithmic development in this field is that the ground truths of hyperspectral images (HSIs) are extremely hard to obtain. Recently a handful of HIC methods are developed based on the graph convolution networks (GCNs), which effectively relieves the scarcity of labeled data for deep learning based HIC methods. To further lift the classification performance, in this work we propose a graph convolution network (GCN) based framework for HSI classification that uses two clustering operations to better exploit multi-hop node correlations and also effectively reduce graph size. In particular, we first cluster the pixels with similar spectral features into a superpixel and build the graph based on the superpixels of the input HSI. Then instead of performing convolution over this superpixel graph, we further partition it into several sub-graphs by pruning the edges with weak weights, so as to strengthen the correlations of nodes with high similarity. This second round of clustering also further reduces the graph size, thus reducing the computation burden of graph convolution. Experimental results on three widely used benchmark datasets well prove the effectiveness of our proposed framework.
Potholes are one of the most common forms of road damage, which can severely affect driving comfort, road safety and vehicle condition. Pothole detection is typically performed by either structural engineers or certified inspectors. This task is, however, not only hazardous for the personnel but also extremely time-consuming. This paper presents an efficient pothole detection algorithm based on road disparity map estimation and segmentation. We first generalize the perspective transformation by incorporating the stereo rig roll angle. The road disparities are then estimated using semi-global matching. A disparity map transformation algorithm is then performed to better distinguish the damaged road areas. Finally, we utilize simple linear iterative clustering to group the transformed disparities into a collection of superpixels. The potholes are then detected by finding the superpixels, whose values are lower than an adaptively determined threshold. The proposed algorithm is implemented on an NVIDIA RTX 2080 Ti GPU in CUDA. The experiments demonstrate the accuracy and efficiency of our proposed road pothole detection algorithm, where an accuracy of 99.6% and an F-score of 89.4% are achieved.
Superpixel segmentation aims at dividing the input image into some representative regions containing pixels with similar and consistent intrinsic properties, without any prior knowledge about the shape and size of each superpixel. In this paper, to alleviate the limitation of superpixel segmentation applied in practical industrial tasks that detailed boundaries are difficult to be kept, we regard each representative region with independent semantic information as a subspace, and correspondingly formulate superpixel segmentation as a subspace clustering problem to preserve more detailed content boundaries. We show that a simple integration of superpixel segmentation with the conventional subspace clustering does not effectively work due to the spatial correlation of the pixels within a superpixel, which may lead to boundary confusion and segmentation error when the correlation is ignored. Consequently, we devise a spatial regularization and propose a novel convex locality-constrained subspace clustering model that is able to constrain the spatial adjacent pixels with similar attributes to be clustered into a superpixel and generate the content-aware superpixels with more detailed boundaries. Finally, the proposed model is solved by an efficient alternating direction method of multipliers (ADMM) solver. Experiments on different standard datasets demonstrate that the proposed method achieves superior performance both quantitatively and qualitatively compared with some state-of-the-art methods.
In this paper, we propose a method called superpixel tensor pooling tracker which can fuse multiple midlevel cues captured by superpixels into sparse pooled tensor features. Our method first adopts the superpixel method to generate different patches (superpixels) from the target template or candidates. Then for each superpixel, it encodes different midlevel cues including HSI color, RGB color, and spatial coordinates into a histogram matrix to construct a new feature space. Next, these matrices are formed to a third order tensor. After that, the tensor is pooled into the sparse representation. Then the incremental positive and negative subspaces learning is performed. Our method has both good characteristics of midlevel cues and sparse representation hence is more robust to large appearance variations and can capture compact and informative appearance of the target object. To validate the proposed method, we compare it with state-of-the-art methods on 24 sequences with multiple visual tracking challenges. Experiment results demonstrate that our method outperforms them significantly.
Most superpixel methods are sensitive to noise and cannot control the superpixel number precisely. To solve these problems, in this paper, we propose a robust superpixel method called fuzzy simple linear iterative clustering (Fuzzy SLIC), which adopts a local spatial fuzzy C-means clustering and dynamic fuzzy superpixels. We develop a fast and precise superpixel number control algorithm called onion peeling (OP) algorithm. Fuzzy SLIC is insensitive to most types of noise, including Gaussian, salt and pepper, and multiplicative noise. The OP algorithm can control the superpixel number accurately without reducing much computational efficiency. In the validation experiments, we tested the Fuzzy SLIC and OP algorithm and compared them with state-of-the-art methods on the BSD500 and Pascal VOC2007 benchmarks. The experiment results show that our methods outperform state-of-the-art techniques in both noise-free and noisy environments.
We present a fast and accurate method for dense depth reconstruction from sparsely sampled light fields obtained using a synchronized camera array. In our method, the source images are over-segmented into non-overlapping compact superpixels that are used as basic data units for depth estimation and refinement. Superpixel representation provides a desirable reduction in the computational cost while preserving the image geometry with respect to the object contours. Each superpixel is modeled as a plane in the image space, allowing depth values to vary smoothly within the superpixel area. Initial depth maps, which are obtained by plane sweeping, are iteratively refined by propagating good correspondences within an image. To ensure the fast convergence of the iterative optimization process, we employ a highly parallel propagation scheme that operates on all the superpixels of all the images at once, making full use of the parallel graphics hardware. A few optimization iterations of the energy function incorporating superpixel-wise smoothness and geometric consistency constraints allows to recover depth with high accuracy in textured and textureless regions as well as areas with occlusions, producing dense globally consistent depth maps. We demonstrate that while the depth reconstruction takes about a second per full high-definition view, the accuracy of the obtained depth maps is comparable with the state-of-the-art results.
Although deep learning greatly improves the performance of semantic segmentation, its success mainly lies in object central areas without accurate edges. As superpixels are a popular and effective auxiliary to preserve object edges, in this paper, we jointly learn semantic segmentation with trainable superpixels. We achieve it with fully-connected layers with Transparent Initialization (TI) and efficient logit consistency using a sparse encoder. The proposed TI preserves the effects of learned parameters of pretrained networks. This avoids a significant increase of the loss of pretrained networks, which otherwise may be caused by inappropriate parameter initialization of the additional layers. Meanwhile, consistent pixel labels in each superpixel are guaranteed by logit consistency. The sparse encoder with sparse matrix operations substantially reduces both the memory requirement and the computational complexity. We demonstrated the superiority of TI over other parameter initialization methods and tested its numerical stability. The effectiveness of our proposal was validated on PASCAL VOC 2012, ADE20K, and PASCAL Context showing enhanced semantic segmentation edges. With quantitative evaluations on segmentation edges using performance ratio and F-measure, our method outperforms the state-of-the-art.
Deep neural networks are becoming more and more popular due to their revolutionary success in diverse areas, such as computer vision, natural language processing, and speech recognition. However, the decision-making processes of these models are generally not interpretable to users. In various domains, such as healthcare, finance, or law, it is critical to know the reasons behind a decision made by an artificial intelligence system. Therefore, several directions for explaining neural models have recently been explored. In this thesis, I investigate two major directions for explaining deep neural networks. The first direction consists of feature-based post-hoc explanatory methods, that is, methods that aim to explain an already trained and fixed model (post-hoc), and that provide explanations in terms of input features, such as tokens for text and superpixels for images (feature-based). The second direction consists of self-explanatory neural models that generate natural language explanations, that is, models that have a built-in module that generates explanations for the predictions of the model.
Segmentation-based tracking has been actively studied in computer vision and multimedia. Superpixel based object segmentation and tracking methods are usually developed for this task. However, they independently perform feature representation and learning of superpixels which may lead to sub-optimal results. In this paper, we propose to utilize graph convolutional network (GCN) model for superpixel based object tracking. The proposed model provides a general end-to-end framework which integrates i) label linear prediction, and ii) structure-aware feature information of each superpixel together to obtain object segmentation and further improves the performance of tracking. The main benefits of the proposed GCN method have two main aspects. First, it provides an effective end-to-end way to exploit both spatial and temporal consistency constraint for target object segmentation. Second, it utilizes a mixed graph convolution module to learn a context-aware and discriminative feature for superpixel representation and labeling. An effective algorithm has been developed to optimize the proposed model. Extensive experiments on five datasets demonstrate that our method obtains better performance against existing alternative methods.
Parkinson's disease (PD) is a degenerative and progressive neurological condition. Early diagnosis can improve treatment for patients and is performed through dopaminergic imaging techniques like the SPECT DaTscan. In this study, we propose a machine learning model that accurately classifies any given DaTscan as having Parkinson's disease or not, in addition to providing a plausible reason for the prediction. This is kind of reasoning is done through the use of visual indicators generated using Local Interpretable Model-Agnostic Explainer (LIME) methods. DaTscans were drawn from the Parkinson's Progression Markers Initiative database and trained on a CNN (VGG16) using transfer learning, yielding an accuracy of 95.2%, a sensitivity of 97.5%, and a specificity of 90.9%. Keeping model interpretability of paramount importance, especially in the healthcare field, this study utilises LIME explanations to distinguish PD from non-PD, using visual superpixels on the DaTscans. It could be concluded that the proposed system, in union with its measured interpretability and accuracy may effectively aid medical workers in the early diagnosis of Parkinson's Disease.
An efficient spatial regularization method using superpixel segmentation and graph Laplacian regularization is proposed for sparse hyperspectral unmixing method. Since it is likely to find spectrally similar pixels in a homogeneous region, we use a superpixel segmentation algorithm to extract the homogeneous regions by considering the image boundaries. We first extract the homogeneous regions, which are called superpixels, then a weighted graph in each superpixel is constructed by selecting $K$-nearest pixels in each superpixel. Each node in the graph represents the spectrum of a pixel and edges connect the similar pixels inside the superpixel. The spatial similarity is investigated using graph Laplacian regularization. Sparsity regularization for abundance matrix is provided using a weighted sparsity promoting norm. Experimental results on simulated and real data sets show the superiority of the proposed algorithm over the well-known algorithms in the literature.
Few-shot semantic segmentation (FSS) has great potential for medical imaging applications. Most of the existing FSS techniques require abundant annotated semantic classes for training. However, these methods may not be applicable for medical images due to the lack of annotations. To address this problem we make several contributions: (1) A novel self-supervised FSS framework for medical images in order to eliminate the requirement for annotations during training. Additionally, superpixel-based pseudo-labels are generated to provide supervision; (2) An adaptive local prototype pooling module plugged into prototypical networks, to solve the common challenging foreground-background imbalance problem in medical image segmentation; (3) We demonstrate the general applicability of the proposed approach for medical images using three different tasks: abdominal organ segmentation for CT and MRI, as well as cardiac segmentation for MRI. Our results show that, for medical image segmentation, the proposed method outperforms conventional FSS methods which require manual annotations for training.
Coronaviruses constitute a family of viruses that gives rise to respiratory diseases. As COVID-19 is highly contagious, early diagnosis of COVID-19 is crucial for an effective treatment strategy. However, the RT-PCR test which is considered to be a gold standard in the diagnosis of COVID-19 suffers from a high false-negative rate. Chest X-ray (CXR) image analysis has emerged as a feasible and effective diagnostic technique towards this objective. In this work, we propose the COVID-19 classification problem as a three-class classification problem to distinguish between COVID-19, normal, and pneumonia classes. We propose a three-stage framework, named COV-ELM. Stage one deals with preprocessing and transformation while stage two deals with feature extraction. These extracted features are passed as an input to the ELM at the third stage, resulting in the identification of COVID-19. The choice of ELM in this work has been motivated by its faster convergence, better generalization capability, and shorter training time in comparison to the conventional gradient-based learning algorithms. As bigger and diverse datasets become available, ELM can be quickly retrained as compared to its gradient-based competitor models. The proposed model achieved a macro average F1-score of 0.95 and the overall sensitivity of ${0.94 \pm 0.02} at a 95% confidence interval. When compared to state-of-the-art machine learning algorithms, the COV-ELM is found to outperform its competitors in this three-class classification scenario. Further, LIME has been integrated with the proposed COV-ELM model to generate annotated CXR images. The annotations are based on the superpixels that have contributed to distinguish between the different classes. It was observed that the superpixels correspond to the regions of the human lungs that are clinically observed in COVID-19 and Pneumonia cases.
This paper tackles the unsupervised depth estimation task in indoor environments. The task is extremely challenging because of the vast areas of non-texture regions in these scenes. These areas could overwhelm the optimization process in the commonly used unsupervised depth estimation framework proposed for outdoor environments. However, even when those regions are masked out, the performance is still unsatisfactory. In this paper, we argue that the poor performance suffers from the non-discriminative point-based matching. To this end, we propose P$^2$Net. We first extract points with large local gradients and adopt patches centered at each point as its representation. Multiview consistency loss is then defined over patches. This operation significantly improves the robustness of the network training. Furthermore, because those textureless regions in indoor scenes (e.g., wall, floor, roof, \etc) usually correspond to planar regions, we propose to leverage superpixels as a plane prior. We enforce the predicted depth to be well fitted by a plane within each superpixel. Extensive experiments on NYUv2 and ScanNet show that our P$^2$Net outperforms existing approaches by a large margin. Code is available at \url{https://github.com/svip-lab/Indoor-SfMLearner}.