A Vehicle Type Recognition Method based on Sparse Auto Encoder

Efficient modeling of high-dimensional data requires extracting only relevant dimensions through feature learning. Unsupervised feature learning has gained tremendous attention due to its unbiased approach, no need for prior knowledge or expensive manual processing, and ability to handle exponential data growth. Deep Autoencoder (AE) is a state-of-the-art deep neural network for unsupervised feature learning, which learns embedded-representations using a series of stacked layers. However, as the AE network gets deeper, these learned embedded-representations can deteriorate due to vanishing gradient, leading to performance degradation. This article presents ResNet Autoencoder (RAE) and its convolutional version (C-RAE) for unsupervised feature learning. The advantage of RAE and C-RAE is that it enables the user to add residual connections for increased network capacity without incurring the cost of degradation for unsupervised feature learning compared to standard AEs. While RAE and C-RAE inherit all the advantages of AEs, such as automated non-linear feature extraction and unsupervised learning, they also allow users to design larger networks without adverse effects on feature learning performance. We performed classification on learned embedded-representation to evaluate RAE and C-RAE. RAE and C-RAE were compared against AEs on MNIST, Fashion MNIST, and CIFAR10 datasets. When increasing the number of layers, C-RAE outperformed AE by showing significantly lower performance degradation of classification accuracy (less than 3%) compared to AE (33% to 65%). Further, C-RAE exhibited higher mean accuracy and lower variance of accuracy than standard AE. When comparing RAE and C-RAE with widely used feature learning methods (Convolutional AE, PCA, ICA, LLE, Factor Analysis, and SVD), C-RAE showed the highest accuracy.

PurposeTo differentiate complicated appendicitis (CA) from uncomplicated appendicitis (UA) in children, we developed and validated an artificial intelligence (AI) model using a multimodal approach integrating ultrasound images and clinical data.MethodsA retrospective analysis was performed on 372 pathologically confirmed pediatric appendicitis cases (230 male, 142 female) from three centers, all with preoperative abdominal ultrasound. Deep learning (DL) features and radiomic features were extracted from appendiceal ultrasound images using deep transfer learning (DTL) and conventional radiomic methods, respectively. We selected 12 radiomic features, 9 DL features, and 3 clinical features—namely, white blood cell count (WBC), neutrophil count (NEU), and C-reactive protein (CRP)—for building the machine learning classification model. Based on these features, four distinct models were constructed: the Rad model (radiomic features only), the DL model (DL features only), the DTL model (combined radiomic and DL features), and the Combine model (integrating all three feature types: radiomic, DL, and clinical features). Model performance was evaluated using receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and the DeLong test. Finally, the combined model was compared to the performance of clinicians with varying levels of experience.ResultsThe combined model demonstrated consistently favorable performance across all cohorts (AUC: 0.940, 0.895, 0.866, and 0.783 for training and validation sets, respectively). The model's accuracy (0.862) and positive predictive value (0.896) were comparable to senior surgeons (0.741, 0.970) and superior to junior surgeons (0.672, 0.865) in the internal validation cohort. DCA confirmed the clinical utility of the combined model over conventional strategies.ConclusionOur ultrasound-based AI model provides reliable differentiation between CA and UA in children, offering potential value as a diagnostic support tool for clinical decision making.

Sleep staging based on electroencephalogram (EEG) signal, as one of the vital bases of study on sleep diagnosis, has been under massive attention. With the spring up of deep learning these years, the idea of combining deep learning structure with automatic sleep staging has been an attractive topic. However, the labeling of sleep stages requires professional knowledge as well as plenty of time, which raise the barrier to evaluate this idea. In this study, the method of unsupervised feature learning based on a mass of unlabeled data and a small number of labeled data was proposed to accomplish sleep staging. The unsupervised feature learning structure was built based on a pair of symmetric convolutional neural networks, with the help of a shallow neural network classifier to classify sleep stages. The results showed that under the condition of the very few labeled data, sleep staging based on unsupervised feature learning can achieve similar accuracy to supervised feature learning, which provides a new direction for the application of deep learning method in dealing with data that is difficult to label or lack of prior knowledge.

In unsupervised learning of image features without labels, especially on datasets with fine-grained object classes, it is often very difficult to tell if a given image belongs to one specific object class or another, even for human eyes. However, we can reliably tell if image C is more similar to image A than image B. In this work, we propose to explore how this relative order can be used to learn discriminative features with an unsupervised metric learning method. Instead of resorting to clustering or self-supervision to create pseudo labels for an absolute decision, which often suffers from high label error rates, we construct reliable relative orders for groups of image samples and learn a deep neural network to predict these relative orders. During training, this relative order prediction network and the feature embedding network are tightly coupled, providing mutual constraints to each other to improve overall metric learning performance in a cooperative manner. During testing, the predicted relative orders are used as constraints to optimize the generated features and refine their feature distance-based image retrieval results using a constrained optimization procedure. Our experimental results demonstrate that the proposed relative orders for unsupervised learning (ROUL) method is able to significantly improve the performance ofunsupervised deep metric learning.

Skeleton-based deep pose feature learning for action quality assessment on figure skating videos

This paper proposes a combination of deep learning and hand-designed feature for plant identification based on leaf and flower images. The contributions of this paper are two-fold. First, for each organ image, we have performed a comparative evaluation of deep learning and hand-designed feature for plant identification. Two approaches for deep learning and hand-designed feature that are convolutional neuron network (CNN) and kernel descriptor (KDES) are chosen in our experiments. Second, based on the results of the first contribution, we propose a method for plant identification by late fusing the identification results of leaf and flower. Experimental results on ImageClef 2015 dataset show that hand designed feature outperforms deep learning for well-constrained cases (leaf captured on simple background). However, deep learning shows its robustness in natural situations. Moreover, the combination of leaf and flower images improves significantly the identification when comparing leaf-based plant identification.

A deep autoencoder feature learning method for process pattern recognition

Sentiment analysis is an important task in the field of big data and artificial intelligence and has a wide range of real-world applications. The majority of current approaches, however, attempt to detect the overall polarity of a sentence, or text span, irrespective of the entities mentioned and their aspects. A typical way for solve this is to employ a classifier that learns to identify the polarity labels given input text. Such methods require amount of training examples which were manually constructed. While most work in text mining in the field of sentiment analysis domain focus on the use of supervised machine learning technologies, in this paper, we investigate on aspect-based sentiment analysis for mobile game reviews. Our method integrates both deep learning learned features and support vector machines. The deep learning feature is mainly derived from inducting word relations in Wikipedia and the collected in-domain text. To prevent from assessing unknown missing words, we design a fuzzy-based approach to capture word-level sentiment scores. By using a fuzzy-based soft computing strategy with a small lexicon allows us to produce a system with better accuracy and precision than pure machine learning such as Naive Bayes and SVMs. To demonstrate the effectiveness of the proposed approach, we conduct the experimental results on the collected Mobile game reviews. These results show that our method outperforms supervised systems. One good property of this method is that it does not need to perform Chinese word segmentation in testing time.

Predicting and Grouping Digitized Paintings by Style using Unsupervised Feature Learning.

BackgroundThe subjective assessment of Breast Imaging Reporting and Data System (BI-RADS) 4A lesions leads to a high number of unnecessary biopsies, highlighting the need for more objective and accurate diagnostic tools. This study aimed to construct and validate a novel multimodal framework that integrates deep learning (DL), radiomics, and clinical and breast ultrasound (US) features to distinguish between benign and malignant breast lesions classified as BI-RADS 4A.MethodsA total of 935 patients with pathologically confirmed BI-RADS 4A lesions were retrospectively enrolled and randomly divided into a training cohort (n=654) and an internal validation cohort (n=281). An additional 488 patients were enrolled as the external validation cohort. DL and radiomics models were developed with the light gradient boosting machine (LightGBM) algorithm, a machine learning technique known for its efficiency in handling high-dimensional data. Univariate and multivariate logistic regression (LR) analyses identified significant clinical and US features. After handcrafted radiomics features, DL features, and clinical/breast US features were combined, dimensionality reduction and feature selection were performed. The integrated model was developed via LightGBM, and the SHapley Additive Explanations (SHAP) approach was applied to rank feature contributions.ResultsThe integrated model achieved robust discrimination between benign and malignant BI-RADS 4A lesions, with an area under the curve (AUC) of 0.938, 0.870, and 0.861 for the training cohort, internal validation cohort, and external validation cohort, respectively. Furthermore, in the external validation cohort, the integrated model significantly outperformed the handcrafted radiomics model (AUC =0.719; P<0.001) and the DL model (AUC =0.763; P=0.011). Decision curve analysis confirmed that the integrated model was the most clinically useful across the three cohorts.ConclusionsThe integrated model demonstrated high accuracy in differentiating between benign and malignant BI-RADS 4A lesions, exhibiting the ability to significantly reduce unnecessary biopsies while maintaining diagnostic accuracy.

Unsupervised feature learning refers to the problem of learning useful feature extraction functions from unlabeled data. Despite the great success of deep learning networks in this task in recent years, both for static and for sequential data, these systems can in general still not compete with the high performance of our brain at learning to extract useful representations from its sensory input. We propose the Neocortex-Inspired Locally Recurrent Neural Network: a new neural network for unsupervised feature learning in sequential data that brings ideas from the structure and function of the neocortex to the well-established fields of machine learning and neural networks. By mimicking connection patterns in the feedforward circuits of the neocortex, our system tries to generalize some of the ideas behind the success of convolutional neural networks to types of data other than images. To evaluate the performance of our system at extracting useful features, we have trained different classifiers using those and other learnt features as input and we have compared the obtained accuracies. Our system has shown to outperform other shallow feature learning systems in this task, both in terms of the accuracies achieved and in terms of how fast the classification task is learnt. The results obtained confirm our system as a state-of-the-art shallow feature learning system for sequential data, and suggest that extending it to or integrating it into deep architectures may lead to new successful networks that are competent at dealing with complex sequential tasks.

Place recognition is one of the major challenges for the LiDAR-based effective localization and mapping task. Traditional methods are usually relying on geometry matching to achieve place recognition, where a global geometry map need to be restored. In this paper, we accomplish the place recognition task based on an end-to-end feature learning framework with the LiDAR inputs. This method consists of two core modules, a dynamic octree mapping module that generates local 2D maps with the consideration of the robot's motion; and an unsupervised place feature learning module which is an improved adversarial feature learning network with additional assistance for the long-term place recognition requirement. More specially, in place feature learning, we present an additional Generative Adversarial Network with a designed Conditional Entropy Reduction module to stabilize the feature learning process in an unsupervised manner. We evaluate the proposed method on the Kitti dataset and North Campus Long-Term LiDAR dataset. Experimental results show that the proposed method outperforms state-of-the-art in place recognition tasks under long-term applications. What's more, the feature size and inference efficiency in the proposed method are applicable in real-time performance on practical robotic platforms.

BackgroundIdentifying predictive markers for immunotherapy in non-small cell lung cancer (NSCLC) is critical for personalized treatment. This study aimed to construct a predictive model that integrates clinical features, enhanced computed tomography (CT)-radiomics, and deep learning (DL) features for the assessment of durable clinical benefit (DCB) from immunotherapy in patients with advanced NSCLC and to provide biological interpretability to predictions by integrating radiogenomic data.MethodsWe conducted a retrospective analysis of 201 advanced NSCLC patients who underwent immunotherapy with CT images, with data supplemented from The Cancer Imaging Archive (TCIA). Radiomics features (RFs) were extracted from enhanced CT images, and DL features were derived using a pre-trained ResNet-34 model. DCB-related signatures were constructed using the least absolute shrinkage and selection operator (LASSO) algorithm, and fusion nomogram models were developed by integrating significant clinical variables, radiomics, and DL features. Shapley additive explanations were employed to quantify the impact of radiomics-DL features on model predictions. Gene set enrichment and biological correlation analyses based on transcriptomic TCIA data were performed to explore the biological significance of radiomics-DL score.ResultsStatistically significant clinical predictors included initial efficacy, brain metastases, programmed death-ligand 1 (PD-L1) expression, and hemoglobin levels. The fusion nomogram model demonstrated the highest predictive accuracy for DCB, with area under the curve (AUC) values of 0.843 in the train cohort and 0.894 in the test cohort, surpassing individual feature sets. Biological exploration revealed associations between radiomics-DL score and biological characteristics, including immune responses and immunoregulation.ConclusionsThis integrated approach shows the potential of combining clinical, radiomics and deep learning features (DLFs) as a noninvasive biomarker for predicting immunotherapy efficacy in NSCLC, assisting in patient selection and clinical decision-making. Radiotranscriptomic analysis may reveal key cellular and immune patterns associated with radiomics-DL signature.

Unsupervised feature learning with deep networks has been widely studied in the recent years. Despite the progress, most existing models would be fragile to non-Gaussian noises and outliers due to the criterion of mean square error (MSE). In this paper, we propose a robust stacked autoencoder (R-SAE) based on maximum correntropy criterion (MCC) to deal with the data containing non-Gaussian noises and outliers. By replacing MSE with MCC, the anti-noise ability of stacked autoencoder is improved. The proposed method is evaluated using the MNIST benchmark dataset. Experimental results show that, compared with the ordinary stacked autoencoder, the R-SAE improves classification accuracy by 14% and reduces the reconstruction error by 39%, which demonstrates that R-SAE is capable of learning robust features on noisy data.

With the rapid development of various satellite sensors, automatic and advanced scene classification technique is urgently needed to process a huge amount of satellite image data. Recently, a few of research works start to implant the sparse coding for feature learning in aerial scene classification. However, these previous research works use the single-layer sparse coding in their system and their performances are highly related with multiple low-level features, such as scale-invariant feature transform (SIFT) and saliency. Motivated by the importance of feature learning through multiple layers, we propose a new unsupervised feature learning approach for scene classification on very high resolution satellite imagery. The proposed unsupervised feature learning utilizes multipath sparse coding architecture in order to capture multiple aspects of discriminative structures within complex satellite scene images. In addition, the dense low-level features are extracted from the raw satellite data by using different image patches with varying size at different layers, and this approach is not limited to a particularly designed feature descriptors compared with the other related works. The proposed technique has been evaluated on two challenging high-resolution datasets, including the UC Merced dataset containing 21 different aerial scene categories with a 1 foot resolution and the Singapore dataset containing 5 land-use categories with a 0.5m spatial resolution. Experimental results show that it outperforms the state-of-the-art that uses the single-layer sparse coding. The major contributions of this proposed technique include (1) a new unsupervised feature learning approach to generate feature representation for very high-resolution satellite imagery, (2) the first multipath sparse coding that is used for scene classification in very high-resolution satellite imagery, (3) a simple low-level feature descriptor instead of many particularly designed low-level descriptor, such as SIFT descriptors and saliency, (4) evaluation on two satellite image datasets that come from different sensor sources.