Variable Stiffness Object Recognition with Bayesian Convolutional Neural Network on a Soft Gripper

Concrete cracks significantly impact structural durability by allowing harmful elements such as carbon dioxide and chloride ions to penetrate. Thus, the cracks should be prevented and treated to maintain the lifespan of concrete structures. Convolutional neural network (CNN) has been widely applied in crack detection. This study investigated the use of a Bayesian convolutional neural network (BCNN) for concrete crack classification and compared the performance of BCNN with UNet, a deep CNN-based model, and CNN. BCNN weight uncertainty through a Bayesian method, improved prediction reliability and prevented overfitting. BCNN demonstrated superior expected calibration error (0.0030) and Brier score (0.0025) compared to CNN, highlighting its robustness in prediction reliability. Furthermore, Bayesian entropy analysis confirmed the ability of BCNN to quantify uncertainty, leading to more precise crack classification. These results imply that BCNN provides a balance between computational efficiency and performance, offering strong potential for smart construction applications. Validation results indicated that BCNN achieved a validation accuracy of 92.22%, higher than CNN (85.78%) and slightly lower than UNet (94.34%). However, BCNN demonstrated a superior intersection over union (IoU) of 0.8049, outperforming CNN (0.6685) and approaching UNet (0.9135). These results implied that BCNN provides a balance between computational efficiency and performance, offering strong potential for smart construction applications.

Soft robotic grippers possess high structural compliance and adaptability, allowing them to grasp objects with unknown and irregular shapes and sizes. To enable more dexterous manipulation, soft sensors that are similar in mechanical properties to common elastomer materials are desired to be integrated into soft grippers. In this paper, we develop ionic hydrogel-based strain and tactile sensors and integrate these sensors into a three-finger soft gripper for learning-based object recognition and force-controlled grasping. Such hydrogel-based sensors have excellent conductivity, high stretchability and toughness, good ambient stability, and unique antifreezing property; they can be readily attached to a soft gripper at desired locations for strain and tactile sensing. By using a deep-learning model, the sensory soft gripper is demonstrated to be capable of grasping and recognizing objects at both room and freezing temperatures, and achieving close to 100% recognition accuracy for ten typical objects. Moreover, the capacitive tactile feedback of the gripper is utilized to develop a closed-loop force controller and realize force-controlled grasping of fragile or highly deformable objects. A new slip detection and compensation strategy is also proposed and validated for the sensory gripper for adjusting the grasping force in real time upon detecting slippage. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —The multimodal sensation of a soft robotic gripper could enrich its grasping functionalities and improve its manipulation performance. This research integrates novel antifreezing ionic hydrogel-based strain and tactile sensors into a three-finger soft robotic gripper for learning-based object recognition and force-controlled grasping. Constructed from a highly stretchable, ambient-stable, and antifreezing ionic hydrogel, the strain and tactile sensors can be readily integrated at the desired locations on the soft gripper, and can reliably operate at both ambient and freezing temperatures with excellent mechanical and electrical properties. Based on the feedback of the strain and tactile sensors, a deep learning model is employed to enable high-accuracy object recognition while grasping, which can be useful for manipulation in vision-free environments. Closed-loop force control and slip compensation strategies are also demonstrated for reliably grasping fragile/deformable objects and handling slip events during the manipulation of heavy objects. The sensory soft gripper and the associated object recognition and force control methods could find practical applications in a variety of robotic manipulation tasks.

Bangla handwritten character recognition is a popular research topic as its difficulty is higher than the recognition of other languages because of multiple formats of compound characters. State of the art Convolutional neural network (CNN) architectures are very much useful in computer vision applications. Some works have been carried out in Bangla handwritten character recognition but most of them either not very efficient or they can not classify a lot of characters. In this work, state of art pre-trained CNN architectures is used to classify 231 different Bangla handwritten characters using CMATERdb dataset. The images were first converted to B&W form with white as the foreground color. The size of the images is reduced to 28 × 28 form. These images are used as input to the CNN architectures. The weights of the state-of-the-art CNN models are kept as it was. The training learning rate was set to 0.001 and categorical cross-entropy as the error function. After 50 epochs, InceptionResNetV2 achieved the best accuracy (96.99%). DenseNet121 and InceptionNetV3 also provided remarkable recognition accuracy (96.55 and 96.20%, respectively). We also considered combination of trained InceptionResNetV2, InceptionNetV3 and DenseNet121 architectures which provided better recognition accuracy (97.69%) than other single CNN architectures but it is not feasible for using as it requires a lot of computation power and memory. The models were tested in the cases where characters look confusing to humans, but all the architectures showed equal capability in recognizing these images. Considering computational complexity, memory and capability of recognizing confused characters, InceptionResNetV2 can be said as the best performing model.

Recognition of human activity based on convolutional neural network has received the interest of researchers in recent years due to its significant improvement in accuracy. A large number of algorithms based on the deep learning approach have been proposed for activity recognition purpose. However, with the increasing advancements in technologies having limited computational resources, it needs to design an efficient deep learning-based approaches with improved utilization of computational resources. This paper presents a simple and efficient 2-Dimensional convolutional neural network (2-D CNN) architecture with very small size convolutional kernel for human activity recognition. The merit of the proposed CNN architecture over standard deep learning architectures is fewer trainable parameters and lesser memory requirement which enables it to train the proposed CNN architecture on low GPU memory-based devices and also works well with smaller as well as larger size datasets. The proposed approach consists of mainly four stages: namely (1) creation of dataset and data augmentation, (2) designing 2-D convolutional neural network (CNN) architecture, (3) the proposed 2-D CNN architecture trained from scratch up to optimum stage, and (4) evaluation of the trained 2-D CNN architecture. To illustrate the effectiveness of the proposed architecture several extensive experiments are conducted on three publicly available datasets, namely IXMAS, YouTube, and UCF101 dataset. The results of the proposed method and its comparison with other state-of-the-art methods [8-12,14,18-26,29-33] demonstrate the usefulness of the proposed method.

In this paper, we study the performance of variants of well-known Convolutional Neural Network (CNN) architectures on different audio tasks. We show that tuning the Receptive Field (RF) of CNNs is crucial to their generalization. An insufficient RF limits the CNN's ability to fit the training data. In contrast, CNNs with an excessive RF tend to over-fit the training data and fail to generalize to unseen testing data. As state-of-the-art CNN architectures - in computer vision and other domains - tend to go deeper in terms of number of layers, their RF size increases and therefore they degrade in performance in several audio classification and tagging tasks. We study well-known CNN architectures and how their building blocks affect their receptive field. We propose several systematic approaches to control the RF of CNNs and systematically test the resulting architectures on different audio classification and tagging tasks and datasets. The experiments show that regularizing the RF of CNNs using our proposed approaches can drastically improve the generalization of models, out-performing complex architectures and pre-trained models on larger datasets. The proposed CNNs achieve state-of-the-art results in multiple tasks, from acoustic scene classification to emotion and theme detection in music to instrument recognition, as demonstrated by top ranks in several pertinent challenges (DCASE, MediaEval).

Every year, new Convolutional Neural Network (CNN) architectures appear to deal with different problems in the activity of image and video recognition. These architectures usually work along the ImageNet dataset for looking for the best performance of the CNNs without taking into account the video task where they are used. This can represent a problem if the task is Human Action Recognition (HAR) in video, since the CNN architectures are pre-trained with an image dataset that can practically contain any object, while HAR problem requires consecutive frames of people doing actions. To prove the idea that using CNNs pre-trained on an image dataset does not always achieve the best performance on a video dataset and that, therefore, it is worth comparing the performance of different CNNs under similar circumstances for the HAR problem, this work proposes an analysis between eight different CNN architectures. Each one of the CNN was exclusively trained with RGB images, which were extracted from the frames of the different classes of videos of HMDB51 dataset. To make the classification of an activity in video, we average the predictions taking into account the successes. We also made some ensembles with the best performance CNNs to measure the improvement in accuracy. Our results suggest that Xception is a strong baseline model that could be used by the community to make their comparisons of their proposals more robust.

Developing non-visual sensing and intelligent recognition technologies is crucial for enhancing the manipulation performance of robots in dim or obstructed environments. Although precise object recognition has been extensively researched for rigid manipulators, the adoption of these techniques in soft robotic systems has been limited by the high modulus or low sensitivity of existing sensors (gauge factor/Young’s modulus<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$&lt;$</tex-math> </inline-formula>10 kPa<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$^{-1}$</tex-math> </inline-formula>). Meanwhile, these systems necessitate advanced artificial intelligence(AI) algorithms to effectively process multiple sensing data. In this study, we utilize newly developed soft sensors and AI algorithm to establish a biomimetic perceptual soft gripper system capable of sensing and generating category object information during the grasping task. A strain/pressure bimodal sensor, mimicking the exceptional softness (Young’s modulus<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$&lt;$</tex-math> </inline-formula>10 kPa) and high sensitivity (gauge factor<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$&gt;$</tex-math> </inline-formula>2000) of human skin, has been developed and seamlessly integrated into a three-finger soft robotic gripper. A Swin Transformer network was developed to learn rules from bimodal data acquired from sensors and generate category information of the grasping objects. The perceptual gripper system exhibited superior recognition accuracy compared to previously reported systems, achieving an impressive 94.9% accuracy in categorizing 18 objects of varying shapes and sizes. We believe this advancement unlocks soft machines’ potential for automated applications. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i>—Non-visual sensing and recognition techniques for soft robotic grippers are valuable for operation in dimly lit or obstructed environments. Traditional rigid sensor have limitations when applied to soft robotic systems due to their high modulus. This research developed novel ultra-soft and high sensitive strain/pressure bimodal sensors integrated into a three-fingered soft robotic gripper. Meanwhile, an AI algorithm applicable to the soft gripper system was developed to process the sensing data and generate the corresponding object category information when grasping the object. With the aid of the developed soft sensor and AI algorithm, the perceptual soft gripper can achieve high-precision object recognition, giving it great potential for automating tasks such as item picking, assembly, and handling in energy-efficient warehouses and logistics centers. The developed soft sensor and AI algorithm co-pilot object recognition strategy can contribute to the advancement of soft machine automation capability technology.

In this paper, the Convolutional Neural Network (CNN) architecture, which is one of the deep learning architectures, is used to classify the basic circuit components drawn by hand. During the training and testing stages of the model, a new dataset containing images of 863 circuit components manually drawn by different people is created. The data set contains images of four different classes of circuit components such as resistor, inductor, capacitor and voltage source. All images have been fixed to the same size and converted to grayscale to increase recognition performance and reduce process complexity. In the study, training for four classes is performed with CNN architecture. Based on the CNN architecture, four new CNN models are employed with different the number of layers. The training and validation results of these models are compared separately, the model with the highest training and validation performance is observed with four layer CNN model (CNN-4). This model obtained 84.41% accuracy rate at classification task.

Inspired by nature's flexible and adaptable organisms, soft robotics are motorless robots made from highly compliant materials to work in confined environments and manipulate delicate objects. However, soft robots often suffer from early failure because of unexpected damage. At the same time, it is challenging to manufacture the geometrically complex structures of soft robots. This study introduces resins based on deep eutectic solvents (DES) to fabricate a pneumatically driven soft gripper using digital light processing (DLP). The resins consist of choline chloride (ChCl) as a hydrogen bond acceptor, glycerol (Gly), and acrylamide (AAm) as hydrogen bond donors. By utilizing the intense hydrogen bonding within DES, the resin can be rapidly cured by photopolymerization to form tough ionogels without chemical crosslinkers. The DES ionogels exhibit remarkable toughness and self‐healing performance compared to common hydrogels. Furthermore, the ionogels show not only efficient energy‐dissipating behavior but also achieve rapid self‐recovery. Finally, the DLP‐printed soft gripper from the DES‐based resin performs successful actuation and healing of macroscopic damages. This work presents a simple strategy to 3D print a soft robotic gripper with high toughness and self‐healing capability.

Classification performance based on ImageNet is the de-facto standard metric for CNN development. In this work we challenge the notion that CNN architecture design solely based on ImageNet leads to generally effective convolutional neural network (CNN) architectures that perform well on a diverse set of datasets and application domains. To this end, we investigate and ultimately improve ImageNet as a basis for deriving such architectures. We conduct an extensive empirical study for which we train 500 CNN architectures, sampled from the broad AnyNetX design space, on ImageNet as well as 8 additional well-known image classification benchmark datasets from a diverse array of application domains. We observe that the performances of the architectures are highlydataset dependent. Some datasets even exhibit a negative error correlation with ImageNet across all architectures. We show how to significantly increase these correlations byutilizing ImageNet subsets restricted to fewer classes. These contributions can have a profound impact on the way we design future CNN architectures and help alleviate the tilt we see currently in our community with respect to over-reliance on one dataset.

Although font recognition is a fundamental issue in the field of document analysis and recognition, it was usually ignored in the past. With the development of optical character recognition (OCR), font recognition becomes more and more important. This paper proposed a well-designed convolutional neural network (CNN) architecture for traditional Mongolian font recognition by means of a single word. To be specific, the whole word image is regarded as input of CNN. Hence, the word images should be normalized into the same size before being inputted into CNN. By comparison, an appropriate aspect ratio for the traditional Mongolian word images has been determined. Experimental results demonstrate that the proposed CNN architecture outperforms three classic CNN architectures, including LeNet-5, AlexNet and GoogLeNet. Therefore, the proposed CNN is much more suitable for the task of the traditional Mongolian font recognition in the way of a single word.

Probabilistic SAR-based water segmentation with adapted Bayesian convolutional neural network

The convolutional neural network (CNN) architecture has shown remarkable success in image classification and segmentation. Its popularity has increased promptly due to various factors such as exponential growth in computational resources, availability of benchmark datasets, supporting libraries, and open-source software. The efficiency of CNN architecture majorly depends on the complexity of the architecture, availability of datasets, and hyperparameter selection. But, due to the huge number of parameters in CNN architecture, its selection has completely remained ad hoc in the past works. In this article, a novel encoding technique has been proposed that can represent complex CNN architecture effectively. The article defines basic building blocks to represent CNN architecture such as the genesis block, transit block, agile block, and output block. This encoding structure is used to generate dynamic length chromosome structure and initialized using evolutionary algorithms. A comparative analysis is also presented that shows compared its effectiveness with existing encoding representations on the basis of the number of encoding parameters, training cost, and efficiency.

Ultrasound (US) is the preferred modality for fatty liver disease diagnosis due to its noninvasive, real-time, and cost-effective imaging capabilities. However, traditional B-mode US is qualitative, and therefore, the assessment is very subjective. Computer-aided diagnostic tools can improve the specificity and sensitivity of US and help clinicians to perform uniform diagnoses. In this work, we propose a novel deep learning model for nonalcoholic fatty liver disease classification from US data. We design a multi-feature guided multi-scale residual convolutional neural network (CNN) architecture to capture features of different receptive fields. B-mode US images are combined with their corresponding local phase filtered images and radial symmetry transformed images as multi-feature inputs for the network. Various fusion strategies are studied to improve prediction accuracy. We evaluate the designed network architectures on B-mode in vivo liver US images collected from 55 subjects. We also provide quantitative results by comparing our proposed multi-feature CNN architecture against traditional CNN designs and machine learning methods. Quantitative results show an average classification accuracy above 90% over tenfold cross-validation. Our proposed method achieves a 97.8% area under the ROC curve (AUC) for the patient-specific leave-one-out cross-validation (LOOCV) evaluation. Comprehensive validation results further demonstrate that our proposed approaches achieve significant improvements compared to training mono-feature CNN architectures ([Formula: see text]). Feature combination is valuable for the traditional classification methods, and the use of multi-scale CNN can improve liver classification accuracy. Based on the promising performance, the proposed method has the potential in practical applications to help radiologists diagnose nonalcoholic fatty liver disease.

Convolutional Neural Network (CNN) architectures for the classification of remote sensing images have been in use for the last 6 years. These CNN architectures are preferred over other standard classifiers, such as SVM and Multilayer perceptron, due to their ability to automatically extract features from image patches. Recently, attention-based models have been used for remote sensing classification because of their ability to concentrate on a few relevant features while ignoring others. To mimic the behavior of the human retina’s ganglia neuron, the concept of On-Off Center-Surround (OOCS) filters has recently been introduced with CNN. This study is an attempt to use both OOCS and attention based deep learning architecture with CNN using two remote sensing datasets. Results suggest that the use of both OOCS and the attention layer does not improve classification accuracy in comparison to the CNN model only.