Search

For large-scale vision tasks in biomedical images, the labeled data is often limited to train effective deep models. Active learning is a common solution, where a query suggestion method selects representative unlabeled samples for annotation, and the new labels are used to improve the base model. However, most query suggestion models optimize their learnable parameters only on the limited labeled data and consequently become less effective for the more challenging unlabeled data. To tackle this, we propose a two-stream active query suggestion approach. In addition to the supervised feature extractor, we introduce an unsupervised one optimized on all raw images to capture diverse image features, which can later be improved by fine-tuning on new labels. As a use case, we build an end-to-end active learning framework with our query suggestion method for 3D synapse detection and mitochondria segmentation in connectomics. With the framework, we curate, to our best knowledge, the largest connectomics dataset with dense synapses and mitochondria annotation. On this new dataset, our method outperforms previous state-of-the-art methods by 3.1% for synapse and 3.8% for mitochondria in terms of region-of-interest proposal accuracy. We also apply our method to image classification, where it outperforms previous approaches on CIFAR-10 under the same limited annotation budget. The project page is https://zudi-lin.github.io/projects/#two_stream_active.

We propose a semi-supervised learning approach for video classification, VideoSSL, using convolutional neural networks (CNN). Like other computer vision tasks, existing supervised video classification methods demand a large amount of labeled data to attain good performance. However, annotation of a large dataset is expensive and time consuming. To minimize the dependence on a large annotated dataset, our proposed semi-supervised method trains from a small number of labeled examples and exploits two regulatory signals from unlabeled data. The first signal is the pseudo-labels of unlabeled examples computed from the confidences of the CNN being trained. The other is the normalized probabilities, as predicted by an image classifier CNN, that captures the information about appearances of the interesting objects in the video. We show that, under the supervision of these guiding signals from unlabeled examples, a video classification CNN can achieve impressive performances utilizing a small fraction of annotated examples on three publicly available datasets: UCF101, HMDB51, and Kinetics.

Most current state-of-the-art connectome reconstruction pipelines have two major steps: initial pixel-based segmentation with affinity prediction and watershed transform, and refined segmentation by merging over-segmented regions. These methods rely only on local context and are typically agnostic to the underlying biology. Since a few merge errors can lead to several incorrectly merged neuronal processes, these algorithms are currently tuned towards over-segmentation producing an overburden of costly proofreading. We propose a third step for connectomics reconstruction pipelines to refine an over-segmentation using both local and global context with an emphasis on adhering to the underlying biology. We first extract a graph from an input segmentation where nodes correspond to segment labels and edges indicate potential split errors in the over-segmentation. In order to increase throughput and allow for large-scale reconstruction, we employ biologically inspired geometric constraints based on neuron morphology to reduce the number of nodes and edges. Next, two neural networks learn these neuronal shapes to further aid the graph construction process. Lastly, we reformulate the region merging problem as a graph partitioning one to leverage global context. We demonstrate the performance of our approach on four real-world connectomics datasets with an average variation of information improvement of 21.3%.

Synaptic connectivity detection is a critical task for neural reconstruction from Electron Microscopy (EM) data. Most of the existing algorithms for synapse detection do not identify the cleft location and direction of connectivity simultaneously. The few methods that computes direction along with contact location have only been demonstrated to work on either dyadic (most common in vertebrate brain) or polyadic (found in fruit fly brain) synapses, but not on both types. In this paper, we present an algorithm to automatically predict the location as well as the direction of both dyadic and polyadic synapses. The proposed algorithm first generates candidate synaptic connections from voxelwise predictions of signed proximity generated by a 3D U-net. A second 3D CNN then prunes the set of candidates to produce the final detection of cleft and connectivity orientation. Experimental results demonstrate that the proposed method outperforms the existing methods for determining synapses in both rodent and fruit fly brain. (Code at: https://github.com/paragt/EMSynConn).

We present an interactive approach to train a deep neural network pixel classifier for the segmentation of neuronal structures. An interactive training scheme reduces the extremely tedious manual annotation task that is typically required for deep networks to perform well on image segmentation problems. Our proposed method employs a feedback loop that captures sparse annotations using a graphical user interface, trains a deep neural network based on recent and past annotations, and displays the prediction output to users in almost real-time. Our implementation of the algorithm also allows multiple users to provide annotations in parallel and receive feedback from the same classifier. Quick feedback on classifier performance in an interactive setting enables users to identify and label examples that are more important than others for segmentation purposes. Our experiments show that an interactively-trained pixel classifier produces better region segmentation results on Electron Microscopy (EM) images than those generated by a network of the same architecture trained offline on exhaustive ground-truth labels.

The prospect of neural reconstruction from Electron Microscopy (EM) images has been elucidated by the automatic segmentation algorithms. Although segmentation algorithms eliminate the necessity of tracing the neurons by hand, significant manual effort is still essential for correcting the mistakes they make. A considerable amount of human labor is also required for annotating groundtruth volumes for training the classifiers of a segmentation framework. It is critically important to diminish the dependence on human interaction in the overall reconstruction system. This study proposes a novel classifier training algorithm for EM segmentation aimed to reduce the amount of manual effort demanded by the groundtruth annotation and error refinement tasks. Instead of using an exhaustive pixel level groundtruth, an active learning algorithm is proposed for sparse labeling of pixel and boundaries of superpixels. Because over-segmentation errors are in general more tolerable and easier to correct than the under-segmentation errors, our algorithm is designed to prioritize minimization of false-merges over false-split mistakes. Our experiments on both 2D and 3D data suggest that the proposed method yields segmentation outputs that are more amenable to neural reconstruction than those of existing methods.

Connectomics has recently begun to image brain tissue at nanometer resolution, which produces petabytes of data. This data must be aligned, labeled, proofread, and formed into graphs, and each step of this process requires visualization for human verification. As such, we present the BUTTERFLY middleware, a scalable platform that can handle massive data for interactive visualization in connectomics. Our platform outputs image and geometry data suitable for hardware-accelerated rendering, and abstracts low-level data wrangling to enable faster development of new visualizations. We demonstrate scalability and extendability with a series of open source Web-based applications for every step of the typical connectomics workflow: data management and storage, informative queries, 2D and 3D visualizations, interactive editing, and graph-based analysis. We report design choices for all developed applications and describe typical scenarios of isolated and combined use in everyday connectomics research. In addition, we measure and optimize rendering throughput—from storage to display—in quantitative experiments. Finally, we share insights, experiences, and recommendations for creating an open source data management and interactive visualization platform for connectomics.

Electron Microscopy (EM) image (or volume) segmentation has become significantly important in recent years as an instrument for connectomics. This paper proposes a novel agglomerative framework for EM segmentation. In particular, given an over-segmented image or volume, we propose a novel framework for accurately clustering regions of the same neuron. Unlike existing agglomerative methods, the proposed context-aware algorithm divides superpixels (over-segmented regions) of different biological entities into different subsets and agglomerates them separately. In addition, this paper describes a "delayed" scheme for agglomerative clustering that postpones some of the merge decisions, pertaining to newly formed bodies, in order to generate a more confident boundary prediction. We report significant improvements attained by the proposed approach in segmentation accuracy over existing standard methods on 2D and 3D datasets.

Pixel and superpixel classifiers have become essential tools for EM segmentation algorithms. Training these classifiers remains a major bottleneck primarily due to the requirement of completely annotating the dataset which is tedious, error-prone and costly. In this paper, we propose an interactive learning scheme for the superpixel classifier for EM segmentation. Our algorithm is ‘active semi-supervised’ because it requests the labels of a small number of examples from user and applies label propagation technique to generate these queries. Using only a small set (< 20%) of all datapoints, the proposed algorithm consistently generates a classifier almost as accurate as that estimated from a complete groundtruth. We provide segmentation results on multiple datasets to show the strength of these classifiers.

We aim to improve segmentation through the use of machine learning tools during region agglomeration. We propose an active learning approach for performing hierarchical agglomerative segmentation from superpixels. Our method combines multiple features at all scales of the agglomerative process, works for data with an arbitrary number of dimensions, and scales to very large datasets. We advocate the use of variation of information to measure segmentation accuracy, particularly in 3D electron microscopy (EM) images of neural tissue, and using this metric demonstrate an improvement over competing algorithms in EM and natural images., 15 pages, 8 figures

We are interested in establishing the correspondence between neuron activity and body curvature during various movements of C. Elegans worms. Given long sequences of images, specifically recorded to glow when the neuron is active, it is required to track all identifiable neurons in each frame. The characteristics of the neuron data, e.g., the uninformative nature of neuron appearance and the sequential ordering of neurons, renders standard single and multi-object tracking methods either ineffective or unnecessary for our task. In this paper, we propose a multi-target tracking algorithm that correctly assigns each neuron to one of several candidate locations in the next frame preserving shape constraint. The results demonstrate how the proposed method can robustly track more neurons than several existing methods in long sequences of images.

Modeling objects using formal grammars has recently regained much attention in computer vision. Probabilistic logic programming, such as Bilattice based Logical Reasoning (BLR), is shown to produce impressive results in object detection/recognition. Although hierarchical object descriptions are preferred in high-level vision tasks for several reasons, BLR has been applied to non-hierarchical object grammars (compositional descriptions of object class). To better align logic programs (esp. BLR) with compositional object hierarchies, we provide a formal grammar, which can guide domain experts to describe objects. That is, we introduce a context-sensitive specification grammar or a meta-grammar, the language of which is the set of all possible object grammars. We show the practicality of the approach by an automatic compiler that translates example object grammars into a BLR logic program and applied it for detecting Graphical User Interface (GUI) components.

This study formulates the IR target detection as a binary classification problem of each pixel. Each pixel is associated with a label which indicates whether it is a target or background pixel. The optimal label set for all the pixels of an image maximizes a posterior distribution of label configuration given the pixel intensities. The posterior probability is factored into (or proportional to) a conditional likelihood of the intensity values and a prior probability of label configuration. Each of these two probabilities are computed assuming a Markov Random Field (MRF) on both pixel intensities and their labels. In particular, this study enforces neighborhood dependency on both intensity values, by a Simultaneous Auto Regressive (SAR) modle, and on labels, by an Auto-Logistic model. The parameters of these MRF models are learned from labeled examples. During testing, an MRF inference technique, namely Iterated Conditional Mode (ICM), produces the optimal label for each pixel. High performances on benchmark datasets demonstrate effectiveness of this method for IR target detection.

The problem we address in this paper is to label datapoints when the information about them is provided primarily in terms of their subsets or groups. The knowledge we have for a group is a numerical weight or likelihood value for each group member to belong to same class. These likelihood values are computed given a class specific model, either explicit or implicit, of the pattern we wish to learn. By defining a Conditional Random Field (CRF) over the labels of data, we formulate the problem as an Markov Network inference problem. We present experimental results for analytical model estimation and object localization where the proposed method produces improved performances over other methods.

The affinity matrix plays the central role in similarity based clustering algorithm. A recent study has shown that, conventional affinity matrices constructed using local neighborhood information are deficient to represent the overall structure of the dataset. In this paper, we propose a novel similarity measure between two points that captures the global setting of the dataset. The proposed affinity measure between two samples is essentially a likelihood that the two samples should fall into the same cluster. To calculate this, we first calculate a pairwise similarity value given a small subset of the data. The distances from a (randomly selected) subset of datapoints to all observations were utilized to produce an intermediate bipartition of the dataset. The outcomes of these bi-partitions provide the subset dependent ‘vote’ in favor of two samples to belong to the same group. These votes are later marginalized to determine the final pairwise similarity values. We achieved better clustering results both synthetic and real images show using affinity matrices learned by proposed voting method than results using the traditional affinity matrices.

Online boosting methods have recently been used successfully for tracking, background subtraction etc. Conventional online boosting algorithms emphasize on interchanging new weak classifiers/features to adapt with the change over time. We are proposing a new online boosting algorithm where the form of the weak classifiers themselves are modified to cope with scene changes. Instead of replacement, the parameters of the weak classifiers are altered in accordance with the new data subset presented to the online boosting process at each time step. Thus we may avoid altogether the issue of how many weak classifiers to be replaced to capture the change in the data or which efficient search algorithm to use for a fast retrieval of weak classifiers. A computationally efficient method has been used in this paper for the adaptation of linear weak classifiers. The proposed algorithm has been implemented to be used both as an online learning and a tracking method. We show quantitative and qualitative results on both UCI datasets and several video sequences to demonstrate improved performance of our algorithm.

Background subtraction is a widely used paradigm to detect moving objects in video taken from a static camera and is used for various important applications such as video surveillance, human motion analysis, etc. Various statistical approaches have been proposed for modeling a given scene background. However, there is no theoretical framework for choosing which features to use to model different regions of the scene background. In this paper we introduce a novel framework for feature selection for background modeling and subtraction. A boosting algorithm, namely RealBoost, is used to choose the best combination of features at each pixel. Given the probability estimates from a pool of features calculated by Kernel Density Estimate (KDE) over a certain time period, the algorithm selects the most useful ones to discriminate foreground objects from the scene background. The results show that the proposed framework successfully selects appropriate features for different parts of the image.