HEAP: Unsupervised Object Discovery and Localization with Contrastive Grouping

Unsupervised object discovery via self-organisation

Unsupervised object discovery (UOD) refers to the task of discriminating the whole region of objects from the background within a scene without relying on labeled datasets, which benefits the task of bounding-box-level localization and pixel-level segmentation. This task is promising due to its ability to discover objects in a generic manner. We roughly categorize existing techniques into two main directions, namely the generative solutions based on image resynthesis, and the clustering methods based on self-supervised models. We have observed that the former heavily relies on the quality of image reconstruction, while the latter shows limitations in effectively modeling semantic correlations. To directly target at object discovery, we focus on the latter approach and propose a novel solution by incorporating weakly-supervised contrastive learning (WCL) to enhance semantic information exploration. We design a semantic-guided self-supervised learning model to extract high-level semantic features from images, which is achieved by fine-tuning the feature encoder of a self-supervised model, namely DINO, via WCL. Subsequently, we introduce Principal Component Analysis (PCA) to localize object regions. The principal projection direction, corresponding to the maximal eigenvalue, serves as an indicator of the object region(s). Extensive experiments on benchmark unsupervised object discovery datasets demonstrate the effectiveness of our proposed solution. The source code and experimental results are publicly available via our project page at https://github.com/npucvr/WSCUOD.git.

A useful capability for a mobile robot is the ability to recognize the objects in its environment that move and change (as distinct from background objects, which are largely stationary). This ability can improve the accuracy and reliability of localization and mapping, enhance the ability of the robot to interact with its environment, and facilitate applications such as inventory management and theft detection. Rather than viewing this task as a difficult application of object recognition methods from computer vision, this work is in line with a recent trend in the community towards unsupervised object discovery and tracking that exploits the fundamentally temporal nature of the data acquired by a robot. Unlike earlier approaches, which relied heavily upon computationally intensive techniques from mapping and computer vision, our approach combines visual features and RGB-D data in a simple and effective way to segment objects from robot sensory data. We then use a Dirichlet process to cluster and recognize objects. The performance of our approach is demonstrated in several test domains.

Unsupervised object discovery aims to localize objects in images, while removing the dependence on annotations required by most deep learning-based methods. To address this problem, we propose a fully unsupervised, bottom-up approach, for multiple objects discovery. The proposed approach is a two-stage framework. First, instances of object parts are segmented by using the intra-image similarity between self-supervised local features. The second step merges and filters the object parts to form complete object instances. The latter is performed by two CNN models that capture semantic information on objects from the entire dataset. We demonstrate that the pseudo-labels generated by our method provide a better precision-recall trade-off than existing single and multiple objects discovery methods. In particular, we provide state-of-the-art results for both unsupervised class-agnostic object detection and unsupervised image segmentation.

Motivated by the recent success of supervised and weakly supervised common object discovery, in this paper, we move forward one step further to tackle common object discovery in a fully unsupervised way. Generally, object co-localization aims at simultaneously localizing objects of the same class across a group of images. Traditional object localization/detection usually trains specific object detectors which require bounding box annotations of object instances, or at least image-level labels to indicate the presence/absence of objects in an image. Given a collection of images without any annotations, our proposed fully unsupervised method is to simultaneously discover images that contain common objects and also localize common objects in corresponding images. Without requiring to know the total number of common objects, we formulate this unsupervised object discovery as a sub-graph mining problem from a weighted graph of object proposals, where nodes correspond to object proposals, and edges represent the similarities between neighbouring proposals. The positive images and common objects are jointly discovered by finding sub-graphs of strongly connected nodes, with each sub-graph capturing one object pattern. The optimization problem can be efficiently solved by our proposed maximal-flow-based algorithm. Instead of assuming that each image contains only one common object, our proposed solution can better address wild images where each image may contain multiple common objects or even no common object. Moreover, our proposed method can be easily tailored to the task of image retrieval in which the nodes correspond to the similarity between query and reference images. Extensive experiments on PASCAL VOC 2007 and Object Discovery data sets demonstrate that even without any supervision, our approach can discover/localize common objects of various classes in the presence of scale, view point, appearance variation, and partial occlusions. We also conduct broad experiments on image retrieval benchmarks, Holidays and Oxford5k data sets, to show that our proposed method, which considers both the similarity between query and reference images and also similarities among reference images, can help to improve the retrieval results significantly.

This paper studies the problem of object discovery - separating objects from the background without manual labels. Existing approaches utilize appearance cues, such as color, texture, and location, to group pixels into object-like regions. However, by relying on appearance alone, these methods fail to separate objects from the background in cluttered scenes. This is a fundamental limitation since the definition of an object is inherently ambiguous and context-dependent. To resolve this ambiguity, we choose to focus on dynamic objects - entities that can move independently in the world. We then scale the recent auto-encoder based frameworks for unsuper-vised object discovery from toy synthetic images to complex real-world scenes. To this end, we simplify their architecture, and augment the resulting model with a weak learning signal from general motion segmentation algorithms. Our experiments demonstrate that, despite only capturing a small subset of the objects that move, this signal is enough to generalize to segment both moving and static instances of dynamic objects. We show that our model scales to a newly collected, photo- realistic synthetic dataset with street driving scenarios. Additionally, we leverage ground truth segmentation and flow annotations in this dataset for thorough ablation and evaluation. Finally, our experiments on the real-world KITTI benchmark demonstrate that the proposed approach outperforms both heuristic- and learning-based methods by capitalizing on motion cues.

Object detection and localization is a paramount important and challenging task in RoboCup MSL (Middle Size League). It has a strong constraint on real-time, as both the robot and obstacles (also robots) are moving quickly. In this paper, a real-time object segmentation approach is proposed, based on a RGB-D camera in which only the range information has been used. The method has four main steps, e.g., point cloud filtering, background points removing, clustering and object localization. Experimental results show that the proposed algorithm can effectively detect and segment objects in 3D space in real-time.

We address an essential problem in computer vision, that of unsupervised foreground object segmentation in video, where a main object of interest in a video sequence should be automatically separated from its background. An efficient solution to this task would enable large-scale video interpretation at a high semantic level in the absence of the costly manual labeling. We propose an efficient unsupervised method for generating foreground object soft masks based on automatic selection and learning from highly probable positive features. We show that such features can be selected efficiently by taking into consideration the spatio-temporal appearance and motion consistency of the object in the video sequence. We also emphasize the role of the contrasting properties between the foreground object and its background. Our model is created over several stages: we start from pixel level analysis and move to descriptors that consider information over groups of pixels combined with efficient motion analysis. We also prove theoretical properties of our unsupervised learning method, which under some mild constraints is guaranteed to learn the correct classifier even in the unsupervised case. We achieve competitive and even state of the art results on the challenging Youtube-Objects and SegTrack datasets, while being at least one order of magnitude faster than the competition. We believe that the strong performance of our method, along with its theoretical properties, constitute a solid step towards solving unsupervised discovery in video.

The conventional object segmentation methods often degrade their performance due to the requirement of user interaction and/or the incomplete colour appearance models. In this Letter, the authors propose a novel design method of a colour appearance model for accurate and fully‐automatic object segmentation by using saliency maps. The authors initialise the Gaussian mixture models (GMMs) to describe the colour appearance of the foreground objects and the background, respectively, where the mean vectors, covariance matrices, and mixing coefficients are updated adaptively such that more salient pixels have larger weights to update the GMM for the foreground objects while less salient pixels have larger weights to update the GMM for the background, respectively. Experiments are performed on MSRC, iCoseg, and PASCAL datasets and we show that the proposed method outperforms the existing methods quantitatively and qualitatively.

Automatically extracting most conspicuous object from an image is useful and important for many computer vision related tasks. Performance of several applications such as object segmentation, image classification based on salient object and content based image editing in computer vision can be improved using this technique. In this research work, performance of structured matrix decomposition with contour based spatial prior is analyzed for extracting salient object from the complex scene. To separate background and salient object, structured matrix decomposition model based on low rank matrix recovery theory is used along with two structural regularizations. Tree structured sparsity inducing regularization is used to capture image structure and to enforce the same object to assign similar saliency values. And, Laplacian regularization is used to enlarge the gap between background part and salient object part. In addition to structured matrix decomposition model, general high level priors along with biologically inspired contour based spatial prior is integrated to improve the performance of saliency related tasks. The performance of the proposed method is evaluated on two demanding datasets, namely, ICOSEG and PASCAL-S for complex scene images. For PASCAL-S dataset precision recall curve of proposed method starts from 0.81 and follows top and right-hand border more than structured matrix decomposition which starts from 0.79. Similarly, structural similarity index score, which is 0.596654 and 0.394864 without using contour based spatial prior and 0.720875 and 0.568001 using contour based spatial prior for ICOSEG and PASCAL-S datasets shows improved result.

Saliency has been widely applied to detecting the most attractive object in an image. This paper presents a novel method based on saliency map for object localization in an image. We regard saliency map computation as a preprocessing step, which is obtained using a discriminative regional feature integration approach. Then, the saliency map is processed to highlight more saliency contour and remove some subordinately salient points using region-grow segmentation algorithm. Finally, saliency object is located with efficient sub-window search algorithm (ESS) in the binarized saliency map. This approach is able to identify the exact location and roughly the size of saliency object in an image. The performance evaluation using MSRA-B dataset demonstrates our approach performs well in object localization in images.

Several formulations based on Random Fields (RFs) have been proposed for joint categorization and segmentation (JCaS) of objects in images. The RF's sites correspond to pixels or superpixels of an image and one defines potential functions (typically over local neighborhoods) which define costs for the different possible assignments of labels to several different sites. Since the segmentation is unknown a priori, one cannot define potential functions over arbitrarily large neighborhoods as that may cross object boundaries. Categorization algorithms extract a set of interest points from the entire image and solve the categorization problem by optimizing cost functions that depend on the feature descriptors extracted from these interest points. There is some disconnect between segmentation algorithms which consider local neighborhoods and categorization algorithms which consider non-local neighborhoods. In this work, we propose to bridge this gap by introducing a novel formulation which uses models of objects with deformable parts, classically used for object categorization, to solve the JCaS problem. We use these models to introduce two new classes of potential functions for JCaS; (a) the first class of potential functions encodes the model score for detecting an object as a function of its visible parts only, and (b) the second class of potential functions encodes shape priors for each visible part and is used to bias the segmentation of the pixels in the support region of the part, towards the foreground object label. We show that most existing deformable parts formulations can be used to define these potential functions and that the resulting potential functions can be optimized exactly using min-cut. As a result, these new potential functions can be integrated with most existing RF-based formulations for JCaS.

<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$360^{\circ }$</tex-math></inline-formula> salient object detection has recently received much attention for 3D scene perception owing to its omnidirectional field of view (FoV). The capability of recognizing salient objects of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$360^{\circ }$</tex-math></inline-formula> images remains technically challenging due to severe spherical distortion. In this paper, we develop a hierarchical uncertainty-aware <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$360^{\circ }$</tex-math></inline-formula> image salient object detection methodology that explicitly explores the geometric and spatial complementary coherence of Tangent projection (TP) and Equirectangular projection (ERP) by a collaborative learning strategy. Concretely, to mitigate spherical distortion, we first intend to learn saliency-related features from less-distorted tangent images, in which a deformation-aware attention block is introduced to mitigate the geometric distortion caused by projecting a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$360^{\circ }$</tex-math></inline-formula> image onto a 2D plane. However, the discrepancies among tangent images pose a new challenge to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$360^{\circ }$</tex-math></inline-formula> image salient object detection. To tackle this issue and achieve accurate localization for salient objects of all sizes, we design a spatial-frequency saliency feature aggregation module to leverage fast Fourier convolution to capture global contextual information from ERP images, such that obtaining more representative saliency features. Moreover, a hierarchical uncertainty-aware bi-projection consistency learning module with strong local-global information embedding capabilities is constructed, which learns the geometric and spatial correlations between tangent images and ERP images via a collaborative learning strategy. Ultimately, salient object maps are produced for <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$360^{\circ }$</tex-math></inline-formula> images on the basis of the merged saliency features driven by the uncertainty. Extensive experiments show that our developed method improves <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${\mathrm{F}}_\beta ^{\sigma }$</tex-math></inline-formula> by an average of 31.67% compared to twenty existing advanced methods on the publicly available 360-SOD dataset.

Fast and accurate salient-object detectors are important for various image processing and computer vision applications, such as adaptive compression and object segmentation. It is also desirable to have a detector that is aware of the position and the size of the salient objects. In this paper, we propose a salient-object detection method that is fast, accurate, and size-aware. For efficient computation, we quantize the image colors and estimate the spatial positions and sizes of the quantized colors. We then feed these values into a statistical model to obtain a probability of saliency. In order to estimate the final saliency, this probability is combined with a global color contrast measure. We test our method on two public datasets and show that our method significantly outperforms the fast state-of-the-art methods. In addition, it has comparable performance and is an order of magnitude faster than the accurate state-of-the-art methods. We exhibit the potential of our algorithm by processing a high-definition video in real time.

Learning feature fusion strategies for various image types to detect salient objects