The spiking neuron model (SNM) mimics the processing paradigm of synaptic and membrane potentials in the cerebral cortex. However, existing SNMs are limited by two issues. First, they lack spike diversity. Although a spiking neuron perceives temporally varying input currents, SNMs only use identical synaptic weights for regulation. Second, they are insensitive to weak spikes. The potential accumulation in SNMs is solely driven by external inputs, ignoring the internal dynamics of potential. Oligodendrocytes, a recent revelation in neuroscience, enhance neural signaling by forming bidirectional communication. This offers the potential to alleviate the aforementioned issues. In this paper, we first propose the mechanism of the oligodendrocyte-spiking neuron (Oli-N) model. Subsequently, using the Oli-N model, we develop our Oli-inspired spiking neural network (Oli-SNN), which broadens the diversity of spike representations and enhances neurons' firing precision through improved sparse coding to enhance weak spikes. Experiments show that our Oli-SNN achieves state-of-the-art performance in the classification task on both static and neuromorphic datasets.

Neural networks derived from the M-P model have excelled in various visual tasks. However, as a simplified simulation version of the brain neural pathway, their structures are locked during training, causing over-fitting and over-parameterization. Although recent models have begun using the biomimetic concept and empirical pruning, they still result in irrational pruning, potentially affecting the accuracy of the model. In this paper, we introduce the Glia unit, composed of oligodendrocytes (Oli) and astrocytes (Ast), to emulate the exact workflow of the mammalian brain, thereby enhancing the biological plausibility of neural functions. Oli selects neurons involved in signal transmission during neural communication and, together with Ast, adaptively optimizes the neural structure. Specifically, we first construct the artificial Glia-Neuron (G-N) model, which is formulated at the instance, group, and interaction levels with adaptive and collaborative mechanisms. Then, we construct GliaNet based on our G-N model, whose structure and connections can be continuously optimized during training. Experiments show that our GliaNet advances state-of-the-art on multiple tasks while significantly reducing its parameters.

Optimization-based meta-learning methods for few-shot one-class classification (FS-OCC) aim to fine-tune a meta-trained model to classify the positive and negative samples using only a few positive samples by adaptation. However, recent approaches primarily focus on adjusting existing meta-learning algorithms for FS-OCC, while overlooking issues stemming from the misalignment between the cross-entropy loss and OCC tasks during adaptation. This misalignment, combined with the limited availability of one-class samples and the restricted diversity of task-specific adaptation, can significantly exacerbate the adverse effects of gradient instability and generalization. To address these challenges, we propose a novel Task-Specific Gradient Adaptation (TSGA) for FS-OCC. Without extra supervision, TSGA learns to generate appropriate, stable gradients by leveraging label prediction and feature representation details of one-class samples and refines the adaptation process by recalibrating task-specific gradients and regularization terms. We evaluate TSGA on three challenging datasets and a real-world CNC Milling Machine application and demonstrate consistent improvements over baseline methods. Furthermore, we illustrate the critical impact of gradient instability and task-agnostic adaptation. Notably, TSGA achieves state-of-the-art results by effectively addressing these issues.

Deep learning models have achieved remarkable success in pathology image analysis. However, they still face challenges in effectively modeling fine-grained, object-level features. Topological Data Analysis (TDA) has shown promise for addressing these issues but remains underexplored, particularly for whole-slide pathology applications. Additionally, the effectiveness of TDA has yet to be firmly established, as current studies largely use small-scale datasets. In this work, we address these gaps by introducing Persistent Homology in Multiple Instance Learning (PMIL), the first adaptable TDA-based module within the MIL framework. We validate our approach on a large-scale classification dataset, benchmarking against multiple state-of-the-art methods.

Early and accurate tumor classification in ultrasound images plays a pivotal role in improving cancer diagnosis and patient outcomes. Existing computer-aided diagnostic (CAD) algorithms often rely on cropping-based single feedforward pathways, which can result in the loss of crucial contextual information around the tumor. The surrounding ultrasound data, including relative intensity, plays a significant role in tumor diagnosis, and incorrect cropping or positioning may lead to unreliable results. To overcome these limitations, we propose a novel Brain-inspired Dual-stream Network (BidsNet), aiming to emulate the functional mechanisms of the dorsal and ventral streams in human visual processing. BidsNet processes the entire ultrasound image as input, preventing errors or loss of contextual details from cropping. The dorsal stream in BidsNet specializes in extracting spatial features, such as shape and texture, while the ventral stream focuses on object recognition and classification. A cross-stream communication mechanism is introduced to facilitate dynamic information sharing between the streams: spatial attention generated in the dorsal stream informs the ventral stream to improve feature localization, while channel attention derived from the ventral stream refines spatial feature representation in the dorsal stream. This collaborative interplay boosts both the interpretability and performance of the network. Extensive experiments on multiple ultrasound datasets demonstrate that BidsNet delivers superior accuracy and interpretability, validating the effectiveness of its dual-stream design and cross-stream communication mechanism.

The artificial neuron (N-N) model-based networks have accomplished extraordinary success for various vision tasks. However, as a simplification of the mammal neuron model, their structure is locked during training, resulting in overfitting and over-parameters. The astrocyte, newly explored by biologists, can adaptively modulate neuronal communication by inserting itself between neurons. The communication, between the astrocyte and neuron, is bidirectionally and shows the potential to alleviate issues raised by unidirectional communication in the N-N model. In this paper, we first elaborate on the artificial Multi-Astrocyte-Neuron (MA-N) model, which enriches the functionality of the artificial neuron model. Our MA-N model is formulated at both astrocyte- and neuron-level that mimics the bidirectional communication with temporal and joint mechanisms. Then, we construct the MA-Net network with the MA-N model, whose neural connections can be continuously and adaptively modulated during training. Experiments show that our MA-Net advances new state-of-the-art on multiple tasks while significantly reducing its parameters by connection optimization.

Network structure learning aims to optimize network architectures and make them more efficient without compromising performance. In this paper, we first study the astrocytes, a new mechanism to regulate connections in the classic M-P neuron. Then, with the astrocytes, we propose an AstroNet that can adaptively optimize neuron connections and therefore achieves structure learning to achieve higher accuracy and efficiency. AstroNet is based on our built Astrocyte-Neuron model, with a temporal regulation mechanism and a global connection mechanism, which is inspired by the bidirectional communication property of astrocytes. With the model, the proposed AstroNet uses a neural network (NN) for performing tasks, and an astrocyte network (AN) to continuously optimize the connections of NN, i.e., assigning weight to the neuron units in the NN adaptively. Experiments on the classification task demonstrate that our AstroNet can efficiently optimize the network structure while achieving state-of-the-art (SOTA) accuracy.

With the concept of teaching being introduced to the machine learning community, a teacher model start using dynamic loss functions to teach the training of a student model. The dynamic intends to set adaptive loss functions to different phases of student model learning. In existing works, the teacher model 1) merely determines the loss function based on the present states of the student model, e.g., disregards the experience of the teacher; 2) only utilizes the states of the student model, e.g., training iteration number and loss/accuracy from training/validation sets, while ignoring the states of the loss function. In this paper, we first formulate the loss adjustment as a temporal task by designing a teacher model with memory units, and, therefore, enables the student learning to be guided by the experience of the teacher model. Then, with a Dynamic Loss Network, we can additionally use the states of the loss to assist the teacher learning in enhancing the interactions between the teacher and the student model. Extensive experiments demonstrate our approach can enhance student learning and improve the performance of various deep models on real-world tasks, including classification, objective detection, and semantic segmentation scenario.

Lung parenchyma segmentation is valuable for improving the performance of lung nodule detection in computed tomography (CT) images. Traditionally, the two tasks are performed separately. This paper proposes a deep multi-task learning (MTL) approach to integrate these tasks for better lung nodule detection. Three new ideas lead to our proposed approach. First, lung parenchyma segmentation is used as the attention module and is combined with nodule detection in a single deep network. Second, lung nodule detection is performed in an anchor-free manner by dividing it into two subtasks, nodule center identification and nodule size regression. Third, a novel pyramid dilated convolution block (PDCB) is proposed to utilize the advantage of dilated convolution and tackle its gridding problem for better lung parenchyma segmentation. Based on these ideas, we design our end-to-end deep network architecture and corresponding MTL method to achieve lung parenchyma segmentation and nodule detection simultaneously. We evaluate the proposed approach on the commonly used Lung Nodule Analysis 2016 (LUNA16) dataset. The experimental results show the value of our contributions and demonstrate that our approach can yield significant improvements compared with state-of-the-art counterparts.

Automatic segmentation of liver tumors in medical images is crucial for computer-aided diagnosis and therapy. It is a challenging task, since the tumors are notoriously small against the background voxels. This paper proposes a new three-stage curriculum learning approach for training deep networks to tackle this small object segmentation problem. The learning in the first stage is performed on the whole input volume to obtain an initial deep network for tumor segmentation. Then the second stage of learning focuses on the tumor-specific features by continuing training the network on the tumor patches. Finally, we retrain the network on the whole input volume in the third stage, in order that the tumor-specific features and the global context can be integrated to improve the final segmentation accuracy. With this approach, we can employ a single network to segment the tumors directly without the need of liver segmentation. We evaluate our approach on a clinical dataset from the hospital and the public MICCAI 2017 Liver Tumor Segmentation (LiTS) Challenge dataset. In the experiments, our approach exhibits significant improvement compared with the commonly used cascade counterpart.

Considering that the global correlation between images is very important for image co-segmentation, we propose a multi-scale Dual-Cross Correlation Network (DCNet) that can efficiently capture global matching information across images to obtain segmentation results. Specifically, the low-dimensional index feature is used to calculate the correlation and the high-dimensional content features are combined with the correlation matrix for final segmentation. Meanwhile, we specially design a Dual-Cross Correlation Module (DCCM) which harvests the spatial and channel correlation with the adjacent pixels of another image on the cross path to enhance the representation of correlation efficiently. By utilizing a further loop operation, each feature can capture the global dependencies from all pixels of another feature. Furthermore, we fuse multi-scale correlation and features into the decoder, which is called Multi-scale Correlation Fusing Decoder (MCFD), to refine the final segmentation results. Moreover, we introduce a new dice loss function to train the whole network by averaging the dice loss value of the foreground and background. Finally, we validate our method on three co-segmentation benchmarks and the results show that our method achieves the state-of-the-art performance.

This paper proposes an effective feature selection method based on monarch butterfly optimization and Fisher criterion. Fisher criterion is applied to evaluate the feature subsets, based on which the optimal feature subsets are searched by using monarch butterfly optimization algorithm. To combine these two components, a method is developed to binarize continuous solution vectors for deciding the feature selection. We conduct experiments on widely used UCI (University of California, Irvine) classification datasets to study the design of our algorithm and compare it with other state-of-the-art counterparts. The experimental results show that the proposed method is reasonable and effective, which achieves the best result of feature selection among the compared methods and has satisfactory efficiency.

We propose a novel discriminative learning approach for Bayesian pattern classification, called ‘constrained maximum margin (CMM)’. We define the margin between two classes as the difference between the minimum decision value for positive samples and the maximum decision value for negative samples. The learning problem is to maximize the margin under the constraint that each training pattern is classified correctly. This nonlinear programming problem is solved using the sequential unconstrained minimization technique. We applied the proposed CMM approach to learn Bayesian classifiers based on Gaussian mixture models, and conducted the experiments on 10 UCI datasets. The performance of our approach was compared with those of the expectation-maximization algorithm, the support vector machine, and other state-of-the-art approaches. The experimental results demonstrated the effectiveness of our approach.

In the big data era, the efficient indexing of gradually increasing databases is becoming vitally important for information retrieval. To incrementally adapt to changes of databases, in this paper we propose a novel clustering based dynamic indexing and retrieval approach. The tree-like indexing st ructure, termed as CD-Tree, updates the structure with constant insertion of data, keeping the tree in consistent with the newest database. The nodes in the CD-Tree are fitted by Gaussian Mixture Models, based on which we design the efficient updating algorithm. The similarity retrieval method utilizing the CD-Tree is further presented, combining one-way search and backtracking strategy to gain good retrieval accuracy and efficiency. We applied the CD-Tree to example-based image retrieval. The experimental results confirm that our approach is effective and promising.

Mixture models are crucial statistical modeling tools at the heart of many challenging applications in computer vision, pattern recognition, and etc. Simplification of mixture models has recently emerged as an important issue in the field of statistical learning. In this paper, we propose a novel Gaussian mixture model simplification approach using only the models parameters, avoiding the use of the original data records which may bring heavy computational and storage burden. We integrate the inter-model similarity and intra-model independence to introduce a similarity measure between two Gaussian mixture models. An objective function is further designed which aims at keeping a balance between model similarity and simplification degree and a heuristic simulated annealing method is presented to search for the optimal parameter set of the simplified model. The experimental results confirm that our approach is effective and promising.

Common CT imaging signs of lung diseases (CISLs) are defined as the imaging signs that frequently appear in lung CT images from patients and play important roles in the diagnosis of lung diseases. This paper proposes a new feature selection method based on FIsher criterion and genetic optimization, called FIG for short, to tackle the CISL recognition problem. In our FIG feature selection method, the Fisher criterion is applied to evaluate feature subsets, based on which a genetic optimization algorithm is developed to find out an optimal feature subset from the candidate features. We use the FIG method to select the features for the CISL recognition from various types of features, including bag-of-visual-words based on the histogram of oriented gradients, the wavelet transform-based features, the local binary pattern, and the CT value histogram. Then, the selected features cooperate with each of five commonly used classifiers including support vector machine (SVM), Bagging (Bag), Naïve Bayes (NB), k -nearest neighbor (k-NN), and AdaBoost (Ada) to classify the regions of interests (ROIs) in lung CT images into the CISL categories. In order to evaluate the proposed feature selection method and CISL recognition approach, we conducted the fivefold cross-validation experiments on a set of 511 ROIs captured from real lung CT images. For all the considered classifiers, our FIG method brought the better recognition performance than not only the full set of original features but also any single type of features. We further compared our FIG method with the feature selection method based on classification accuracy rate and genetic optimization (ARG). The advantages on computation effectiveness and efficiency of FIG over ARG are shown through experiments.

This paper proposes an approach to recognize Common Imaging Signs of Lesions (CISLs) in lung CT images. We combine the bag-of-visual-words based on the Histograms of Oriented Gradients (HOG) and the Local Binary Pattern (LBP) to represent regions of interest (ROIs) in lung CT images. Then the Max-Min posterior Pseudo-probabilities (MMP) learning method is applied to recognize the category of the imaging sign contained in each ROI. We conducted the 5-fold cross validation experiments on a set of 696 ROIs captured from real lung CT images. The proposed approach achieved the average sensitivity of 91.8%, the average specificity of 98.5% and the average accuracy of 98%. Furthermore, the HOG-LBP features surpassed individual HOG or LBP as well as the hybrid of LBP and intensity histograms, and the MMP behaved better than the Support Vector Machines (SVMs). These experimental results confirm the effectiveness of our approach.

In this paper, we first present a hybrid optimization method of Covariance Matrix Adaptation Evolution Strategy, (called MDDS-C-CMA-ES), which is based on Cholesky decomposition (Cholesky-CMA-ES) and local descent search with mean direction. Then we design a parallel version of the method based on the GPU with C-CUDA to solve the problem of large dimensionality. The main advantage of the MDDS-C-CMA-ES method is that every individual is locally searched with the direction that point to the mean vector before being used to calculate a new mean vector. The algorithm can effectively accelerate the convergence speed of the CMA-ES. And the parallel algorithm on the GPU can significantly reduce the computation time further. In order to test the performance we present two experiments. First, we use the serial algorithm to optimize some classical benchmark functions. The results show our method has better performance than NES[2]and CMA-ES. Then we test the performance of the parallel version on some benchmark functions with 1000-dimension and 1500-dimension. The results show the algorithm obtains a 68x speedup in case of 1000-dimension and 90× speedup in case of 1500 dimension.

The optimization method is one of key issues in discriminative learning of pattern classifiers. This paper proposes a hybrid approach of the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) and the gradient decent method for optimizing Bayesian classifiers under the SOFT target based Max-Min posterior Pseudo-probabilities (Soft-MMP) learning framework. In our hybrid optimization approach, the weighted mean of the parent population in the CMA-ES is adjusted by exploiting the gradient information of objective function, based on which the offspring is generated. As a result, the efficiency and the effectiveness of the CMA-ES are improved. We apply the Soft-MMP with the proposed hybrid optimization approach to handwritten digit recognition. The experiments on the CENPARMI database show that our handwritten digit classifier outperforms other state-of-the-art techniques. Furthermore, our hybrid optimization approach behaved better than not only the single gradient decent method but also the single CMA-ES in the experiments.

The problem of determining the appropriate number of components is important in finite mixture modeling for pattern classification. This paper considers the application of an unsupervised clustering method called AutoClass to training of orthogonal Gaussian mixture models (OGMM). Actually, the number of components in OGMM of each class is selected based on AutoClass. In this way, the structures of OGMM for difference classes are not necessarily be the same as those in usual modeling scheme, so that the dissimilarity between the data distributions of different classes can be described more exactly. After the model selection is completed, a discriminative learning framework of Bayesian classifiers called max-min posterior pseudo-probabilities (MMP) is employed to estimate component parameters in OGMM of each class. We apply the proposed learning approach of OGMM to handwritten digit recognition. The experimental results on the MNIST database show the effectiveness of our approach.

This paper proposes a recognition based approach to handwritten numeral string segmentation. We consider two classes: numeral strings segmented correctly or not. The feature vectors containing recognition information for numeral strings segmented correctly are assumed to be of the distribution of Gaussian mixture model (GMM). Based on this modeling, the recognition based segmentation is solved under the max-min posterior pseudo-probabilities (MMP) framework of learning Bayesian classifiers. In the training phase, we use the MMP method to learn a posterior pseudo-probability measure function from positive samples and negative samples of numeral strings segmented correctly. In the process of recognition based segmentation, we generate all possible candidate segmentations of an input string through contour and profile analysis, and then compute the posterior pseudo-probabilities of being the numeral string segmented correctly for all the candidate segmentations. The candidate segmentation with the maximum posterior pseudo-probability is taken as the final result. The effectiveness of our approach is demonstrated by the experiments of numeral string segmentation and recognition on the NIST SD19 database.

Semantic gap is the main problem in current content-based image retrieval. This paper proposes an approach which aims to learn semantic concepts from visual features. Each concept is modeled as a posterior pseudo-probability function, and the function parameters are trained from the positive and negative image examples of the concept using the max-min posterior pseudo-probabilities criterion. According to the posterior pseudo-probabilities of the query concept for all images, the image retrieval is realized by classifying all images into two categories: relevant to the query concept and irrelevant. The number of relevant images can be determined automatically. We show the effectiveness and the advantage of our approach through the experiments on Corel database.

This paper proposes a new method to extract multilingual text in images through discriminating characters from non-characters based on the Gaussian mixture modeling of neighbor characters. The image is binarized and the morphological closing operation is performed on the binary image, in order that each character in it can be treated as a connected component; the neighborhood of connected components are computed based on the Voronoi partition of the image, and each connected component is labeled as character or non-character according to its neighbors. We applied the proposed text extraction method to Chinese and English text extraction, the effectiveness of which is confirmed by the experimental results.

This paper proposes a novel text input method based on hand gestures for wearable computers. A character is firstly written by a fingertip and then recognized through a B-splines based character recognition method. The writing procedure is controlled by hand gestures including hand tracking, gesture recognition and fingertip positioning which are performed by an extended CONDENSATION algorithm. We have integrated the proposed text input method into a wearable vision system developed in our lab and tested the resulting text input system on Graffiti 2 alphabet. The experimental result shows that our method is promising for natural text input in wearable computers.