Accurate Object Recognition Research Articles

Comparison of CNN-Based Architectures for Detection of Different Object Classes

(1) Background: Detecting people and technical objects in various situations, such as natural disasters and warfare, is critical to search and rescue operations and the safety of civilians. A fast and accurate detection of people and equipment can significantly increase the effectiveness of search and rescue missions and provide timely assistance to people. Computer vision and deep learning technologies play a key role in detecting the required objects due to their ability to analyze big volumes of visual data in real-time. (2) Methods: The performance of the neural networks such as You Only Look Once (YOLO) v4-v8, Faster R-CNN, Single Shot MultiBox Detector (SSD), and EfficientDet has been analyzed using COCO2017, SARD, SeaDronesSee, and VisDrone2019 datasets. The main metrics for comparison were mAP, Precision, Recall, F1-Score, and the ability of the neural network to work in real-time. (3) Results: The most important metrics for evaluating the efficiency and performance of models for a given task are accuracy (mAP), F1-Score, and processing speed (FPS). These metrics allow us to evaluate both the accuracy of object recognition and the ability to use the models in real-world environments where high processing speed is important. (4) Conclusion: Although different neural networks perform better on certain types of metrics, YOLO outperforms them on all metrics, showing the best results of mAP-0.88, F1-0.88, and FPS-48, so the focus was on these models.

Deep learning-assisted object recognition with hybrid triboelectric-capacitive tactile sensor

Tactile sensors play a critical role in robotic intelligence and human-machine interaction. In this manuscript, we propose a hybrid tactile sensor by integrating a triboelectric sensing unit and a capacitive sensing unit based on porous PDMS. The triboelectric sensing unit is sensitive to the surface material and texture of the grasped objects, while the capacitive sensing unit responds to the object’s hardness. By combining signals from the two sensing units, tactile object recognition can be achieved among not only different objects but also the same object in different states. In addition, both the triboelectric layer and the capacitor dielectric layer were fabricated through the same manufacturing process. Furthermore, deep learning was employed to assist the tactile sensor in accurate object recognition. As a demonstration, the identification of 12 samples was implemented using this hybrid tactile sensor, and an recognition accuracy of 98.46% was achieved. Overall, the proposed hybrid tactile sensor has shown great potential in robotic perception and tactile intelligence.

MSSA: Multi-Representation Semantics-Augmented Set Abstraction for 3D Object Detection

Accurate recognition and localization of 3D objects is a fundamental research problem in 3D computer vision. Benefiting from transformation-free point cloud processing and flexible receptive fields, point-based methods have become accurate in 3D point cloud modeling, but still fall behind voxel-based competitors in 3D detection. We observe that the set abstraction module, commonly utilized by point-based methods for downsampling points, tends to retain excessive irrelevant background information, thus hindering the effective learning of features for object detection tasks. To address this issue, we propose MSSA, a Multi-representation Semantics-augmented Set Abstraction for 3D object detection. Specifically, we first design a backbone network to encode different representation features of point clouds, which extracts point-wise features through PointNet to preserve fine-grained geometric structure features, and adopts VoxelNet to extract voxel features and BEV features to enhance the semantic features of key points. Second, to efficiently fuse different representation features of keypoints, we propose a Point feature-guided Voxel feature and BEV feature fusion (PVB-Fusion) module to adaptively fuse multi-representation features and remove noise. At last, a novel Multi-representation Semantic-guided Farthest Point Sampling (MS-FPS) algorithm is designed to help set abstraction modules progressively downsample point clouds, thereby improving instance recall and detection performance with more important foreground points. We evaluate MSSA on the widely used KITTI dataset and the more challenging nuScenes dataset. Experimental results show that compared to PointRCNN, our method improves the AP of “moderate” level for three classes of objects by 7.02%, 6.76%, and 5.44%, respectively. Compared to the advanced point-voxel-based method PV-RCNN, our method improves the AP of “moderate” level by 1.23%, 2.84%, and 0.55% for the three classes, respectively.

Application and integration of computer vision technologies for automated recognition and recording of chemical experiments

Abstract Advancements in computer vision have significantly improved motion and object recognition accuracy. These advancements should aid the automatic recognition of chemical experiments, potentially contributing to the recording of experiments. Creating an electronic laboratory notebook from experiment filming enhances convenience and allows more detailed information storage compared to traditional manual recording methods. Our previous research focused on employing object detection and action recognition to automate the recognition of chemical experiments. This paper presents a novel system that combines object detection, action recognition, multiple object tracking, and barcode recognition to automatically generate experimental flowcharts. We implemented our system as a graphical user interface-based application for laboratory use that successfully constructs flowcharts from videos of chemical experiments, including simple chemical manipulations.

A Real-Time Semantic Map Production System for Indoor Robot Navigation.

Although grid maps help mobile robots navigate in indoor environments, some lack semantic information that would allow the robot to perform advanced autonomous tasks. In this paper, a semantic map production system is proposed to facilitate indoor mobile robot navigation tasks. The developed system is based on the employment of LiDAR technology and a vision-based system to obtain a semantic map with rich information, and it has been validated using the robot operating system (ROS) and you only look once (YOLO) v3 object detection model in simulation experiments conducted in indoor environments, adopting low-cost, -size, and -memory computers for increased accessibility. The obtained results are efficient in terms of object recognition accuracy, object localization error, and semantic map production precision, with an average map construction accuracy of 78.86%.

Research on abnormal object detection network of computer room inspection robot based on depth vision

Abstract This paper investigates a deep learning-based anomaly object detection network for identifying and alerting on abnormal items within computer room. First, the framework of the data center inspection robot system is outlined, and the anomaly detection task is decomposed. Next, a dataset of abnormal objects based on data center environmental information is established, and augmentation operations are performed on the created dataset. Subsequently, a SqueezeNet network model based on Residual Squeeze Excitation and Atrous Spatial Pyramid Pooling (RSE-ASPP) is proposed to optimize and improve the SqueezeNet network model. Finally, this paper employs transfer learning to address the issue of insufficient data volume. By pre-training on a large-scale dataset and fine-tuning on the constructed dataset, the accuracy and stability of abnormal object recognition can be significantly enhanced. Ultimately, the proposed RSE-ASPP-SqueezeNet network achieves high-precision detection of abnormal items in the data center inspection robot's anomaly detection task.

MACHINE LEARNING ALGORITHM FOR RAPID OBJECT DETECTION BASED ON COLOR FEATURES

The identification of objects is of utmost importance in a wide range of computer vision applications, such as surveillance systems, autonomous cars, and environmental monitoring. Accurate and efficient object recognition methods are crucial in pastoral environments, characterized by the prominent presence of cattle and other objects, to provide effective analysis and decision-making processes. The present study paper introduces an innovative methodology for efficient identification of objects in pastoral landscapes through the utilization of a Colour Feature Extreme Learning Machine (CF-ELM). The CF-ELM method integrates color characteristics with the ELM algorithm to attain enhanced object detection accuracy while preserving computational economy. The experimental findings provide empirical evidence supporting the efficacy and efficiency of the suggested approach in the detection of items within pastoral landscapes. In addition to the CF-ELM, an algorithm for desktop-based categorization of items within a pastoral environment is provided, with individual speeds ranging from 0.05 s to 0.17 s for a single image, evaluated in each color space. The algorithm is intended for usage in scenarios with challenging and variable terrain, making it appropriate for application in agricultural or pastoral settings.

A high-precision and efficient method for badminton action detection in sports using You Only Look Once with Hourglass Network

Standardized striking movements are essential in badminton for enhancing player techniques and minimizing sports-related injuries. However, accurately detecting these movements against complex backgrounds while balancing precision and speed remains a significant challenge. To address this, we propose a novel model that synergizes You Only Look Once (YOLO) with the Hourglass Network (HGNet), called YOLO-HGNet, to enhance feature learning across multiple levels. By replacing traditional convolutional modules with Depth-Wise Convolution (DWConv), we achieve significant improvements in data processing efficiency. Additionally, our model incorporates a combination of self-attention and convolution mechanisms (ACmix) and FocalModulation to improve object localization and recognition accuracy in complex backgrounds. Our method leverages action vectors and machine learning techniques to accurately detect and classify six key badminton strokes: backhand push, backhand net shot, forehand clear, forehand push, forehand lift, and forehand net shot. Empirical evaluations demonstrate that our approach achieves a mean Average Precision (mAP) of 96.1% for detecting badminton player postures, outperforming existing advanced methods by at least 8.8%. Furthermore, our method achieves an average accuracy of 95.4% in classifying the six badminton strokes. These results underscore the superior capability of YOLO-HGNet for precise and efficient pose detection, recognition, and classification of badminton strokes, contributing significantly to advancements in sports science and athlete training methodologies.

Multimodal tactile sensing fused with vision for dexterous robotic housekeeping

As robots are increasingly participating in our daily lives, the quests to mimic human abilities have driven the advancements of robotic multimodal senses. However, current perceptual technologies still unsatisfied robotic needs for home tasks/environments, particularly facing great challenges in multisensory integration and fusion, rapid response capability, and highly sensitive perception. Here, we report a flexible tactile sensor utilizing thin-film thermistors to implement multimodal perceptions of pressure, temperature, matter thermal property, texture, and slippage. Notably, the tactile sensor is endowed with an ultrasensitive (0.05 mm/s) and ultrafast (4 ms) slip sensing that is indispensable for dexterous and reliable grasping control to avoid crushing fragile objects or dropping slippery objects. We further propose and develop a robotic tactile-visual fusion architecture that seamlessly encompasses multimodal sensations from the bottom level to robotic decision-making at the top level. A series of intelligent grasping strategies with rapid slip feedback control and a tactile-visual fusion recognition strategy ensure dexterous robotic grasping and accurate recognition of daily objects, handling various challenging tasks, for instance grabbing a paper cup containing liquid. Furthermore, we showcase a robotic desktop-cleaning task, the robot autonomously accomplishes multi-item sorting and cleaning desktop, demonstrating its promising potential for smart housekeeping.

Synthetic lidar point cloud generation using deep generative models for improved driving scene object recognition

The imbalanced distribution of different object categories poses a challenge for training accurate object recognition models in driving scenes. Supervised machine learning models trained on imbalanced data are biased and easily overfit the majority classes, such as vehicles and pedestrians, which appear more frequently in driving scenes. We propose a novel data augmentation approach for object recognition in lidar point cloud of driving scenes, which leverages probabilistic generative models to produce synthetic point clouds for the minority classes and complement the original imbalanced dataset. We evaluate five generative models based on different statistical principles, including Gaussian mixture model, variational autoencoder, generative adversarial network, adversarial autoencoder and the diffusion model. Experiments with a real-world autonomous driving dataset show that the synthetic point clouds generated for the minority classes by the Latent Generative Adversarial Network result in significant improvement of object recognition performance for both minority and majority classes. The codes are available at https://github.com/AAAALEX-XIANG/Synthetic-Lidar-Generation.

Dual-mode temperature sensor based on ferroelectric Bi0.5Na0.5TiO3 materials for robotic tactile perception

Temperature perception can allow robots to acquire rich information regarding their surroundings to improve the accuracy of both object manipulation and object recognition. However, existing robotic temperature perception systems mainly depend on thermoresistive effect, suffering from singular functionality and high energy consumption. Here, we demonstrate a ferroelectric Bi0.5Na0.5TiO3-based dual-mode temperature sensor for robotic temperature perception. The dual-mode temperature sensor can operate in a passive mode to perceive temperature variations by generating pyroelectric signals, and work in an active mode to detect static temperatures by producing pyro-photoelectric signals. Equipped with the dual-mode temperature sensor, a robot hand can perceive multi-dimensional temperature information and assess the thermal conductivity of materials, allowing identification of different materials in real-time with a high accuracy of 97.2 %. By combining passive and active temperature perception, the proposed dual-mode temperature sensor therefore holds significant potential for future robotic perception and manipulation.

A novel deep learning method for motion assessment in upper limb rehabilitation grasping test

Purpose The accuracy and reliability of upper limb motion assessment have received great attention in the field of rehabilitation. Grasping test is widely carried out for motion assessment, which requires patients to grasp objects and move them to target place. The traditional assessments test the upper limb motion ability by therapists, which mainly relies on experience and lacks quantitative indicators. This paper aims to propose a deep learning method based on the vision system of our upper limb rehabilitation robot to recognize the motion trajectory of rehabilitation target objects automatically and quantitatively assess the upper limb motion in the grasping test. Design/methodology/approach To begin with, an SRF network is designed to recognize rehabilitation target objects grasped in assessment tests. Moreover, the upper limb motion trajectory is calculated through the motion of objects’ central positions. After that, a GAE network is designed to analyze the motion trajectory which reflects the motion of upper limb. Finally, based on the upper limb rehabilitation exoskeleton platform, the upper limb motion assessment tests are carried out to show the accuracy of both object recognition of SRF network and motion assessment of GAE network. The results including object recognition, trajectory calculation and deviation assessment are given with details. Findings The performance of the proposed networks is validated by experiments that are developed on the upper limb rehabilitation robot. It is implemented by recognizing rehabilitation target objects, calculating the motion trajectory and grading the upper limb motion performance. It illustrates that the networks, including both object recognition and trajectory evaluation, can grade the upper limb motion functionn accurately, where the accuracy is above 95.0% in different grasping tests. Originality/value A novel assessment method of upper limb motion is proposed and verified. According to the experimental results, the accuracy can be remarkably enhanced, and the stability of the results can be improved, which provide more quantitative indicators for further application of upper limb motion assessment.

An efficient Transformer with neighborhood contrastive tokenization for hyperspectral images classification

The success of vision Transformers (ViTs) relies heavily on the self-attention mechanism, which requires support from appropriate patch tokenization. However, hyperspectral image (HSI) often suffer from significant noise distortions and spectral uncertainty, which result in unstable attention patterns and overfitting due to equivocal tokenization. In this paper, we propose a neighborhood contrastive tokenization task (NeiCoT) to learn compact, semantically meaningful, and context-sensitive tokens for efficient Transformer encoding. Specifically, we employ a predictor on patch embedding to maximize the mutual information between local individuals and their global average anchor. This encourages neighboring tokens’ relevance and active participation in feature learning. Next, we revise a token-level contrastive loss to align predictions with local individuals and distinguish them from other samples in a mini-batch to enhance tokens rich in contextual semantics. Furthermore, we apply a Gaussian weighting to the tokens’ contrastive loss to balance the neighborhood contribution. Finally, we propose a sequence-specific MAE framework with NeiCoT to achieve HSI representation, and additionally validate NeiCoT on a supervised Transformer backbone. The results demonstrate that NeiCoT consistently enhances the robustness and generalization of the Transformer, achieving accurate object recognition and boundary localization even with limited training samples. Our code will be available at https://github.com/zoegnov07/NeiCoT.

Image Detection Network Based on Enhanced Small Target Recognition Details and Its Application in Fine Granularity

Image detection technology is of paramount importance across various fields. This significance is not only seen in general images with everyday scenes but also holds substantial research value in the field of remote sensing. Remote sensing images involve capturing images from aircraft or satellites. These images typically feature diverse scenes, large image formats, and varying imaging heights, thus leading to numerous small-sized targets in the captured images. Accurately identifying these small targets, which may occupy only a few pixels, is a challenging and active research area. Current methods mainly fall into two categories: enhancing small target features by improving resolution and increasing the number of small targets to bolster training datasets. However, these approaches often fail to address the core distinguishing features of small targets in the original images, thus resulting in suboptimal performance in fine-grained classification tasks. To address this situation, we propose a new network structure DDU (Downsample Difference Upsample), which is based on differential and resolution changing methods in the Neck layer of deep learning networks to enhance the recognition features of small targets, thus further improving the feature richness of recognition and effectively solving the problem of low accuracy in small target object recognition. At the same time, in order to take into account the recognition effect of targets of other sizes in the image, a new attention mechanism called PNOC (protecting the number of channels) is proposed, which integrates small target features and universal object features without losing the number of channels, thereby increasing the accuracy of recognition. And experimental verification was conducted on the PASCAL-VOC dataset. At the same time, it was applied to the testing of the fine-grained MAR20 dataset and found that the performance was better than other classic algorithms. At the same time, because the proposed framework belongs to a one-stage detection method, it has good engineering applicability and scalability, and universality in scientific research applications are good. Through comparative experiments, it was found that our algorithm improved the performance of the mAP by 0.7% compared to the original YOLOv8 algorithm.

DEEP LEARNING FOR AUTOMATIC CLASSIFICATION OF AVOCADO FRUIT MATURITY

Avocado (Persea Americana), a fleshy fruit with a single seed, has increased in popularity globally, especially in tropical and Mediterranean climates, thanks to its commercial and nutritional value. Rich in bioactive compounds, avocados contribute to the prevention and treatment of various diseases, including cardiovascular problems and cancer. Avocado production in Indonesia, for example, is showing a significant increase, reflecting the growing demand. Avocado ripeness affects shelf life and quality, making the determination of ripeness level a critical aspect of postharvest management. Skin color and pulp firmness change during storage, affecting quality and nutritional value. Proper classification of ripeness is important to reduce post-harvest losses, improve quality and optimize export costs. Recent research shows the use of technologies such as machine learning and YOLO (You Only Look Once) version 9 in real-time detection of avocado ripeness, offering innovative solutions to reduce post-harvest losses and improve distribution efficiency. This approach not only benefits farmers and consumers but also ensures consumer satisfaction and reduces economic losses. This study highlights the importance of real-time detection in monitoring avocado ripeness, where the training process was conducted for 89,280 iterations resulting in a new model for avocado ripeness detection. The final model has a mean Average Precision (mAP) validation value of 84.3%, mAP 84.3% signifies the optimal level of accuracy in object recognition in avocado fruit maturity images using the YOLO model that has undergone an intensive training process.

An Intelligent Apple Identification Method via the Collaboration of YOLOv5 Algorithm and Fast-Guided Filter Theory

Apple-picking robot can promote the development of smart agriculture, and accurate object recognition in complex natural environments using deep learning algorithms is critical. However, research has shown that changes in illumination and object occlusion remain significant challenges for recognition. In order to improve the accuracy of apple apple-picking robot’s identification and positioning of apples in natural environment, a method using YOLOv5 (You Only Look Once, YOLO) combined with fast-guided filter is proposed. By introducing a fast-guided filtering module, the ability to extract image features is improved, and the problem of inaccurate occlusion targets and edge detection is solved; [Formula: see text]-means clustering algorithm is introduced in improving YOLOv5, which can realize automatic adjustment of image size and step size; BiFPN structure is introduced in Neck network to add weighted feature fusion to highlight the detailed features. The results show that the algorithm proposed in this paper can well remove noise information such as occlusion edge blurring in apple images in a natural light environment. In the real orchard environment, the apple recognition accuracy rate reached 97.8%, the recall rate was 97.3% and the recognition rate was about 26.84[Formula: see text]fps. The results show that this research based on YOLOv5 and fast-guided filtering can realize fast and accurate identification of apple fruits in natural environment, and meet the practical application requirements of real-time target detection.

Advanced method for Image detection using OpenCV

Abstract: In computer vision, object detection is an essential job with many real-world applications, such as robots, autonomous cars, and surveillance. Because of its great accuracy and speed, the YOLO (You Only Look Once) method is a well-liked realtime object recognition technique that has attracted a lot of attention. This technique is perfect for time-sensitive applications since it examines the full image at once and predicts bounding boxes and class probabilities for recognized items. YOLO has undergone many iterations, with YOLO v5 being the most recent and sophisticated version. It utilizes anchor boxes and a feature pyramid network (FPN) to increase the accuracy of object recognition. Our goal in this research is to apply YOLO v5 to realtime picture and object identification applications stage. The model will be trained using an appropriate dataset, and its performance will be assessed against a range of benchmarks and advanced object detection methods. The project's output will offer a reliable and effective real-time object detection system that will facilitate prompt decision-making in the identification of item types and their corresponding placements. It has useful uses in robotics, autonomous driving, and surveillance.

Al Based Automatic alarm Generation and Dropping of Payload at a Particular Object through a Drone

Unmanned aerial vehicles (UAVs), commonly referred to as drones, are becoming more and more prevalent in a number of industries, including logistics, surveillance, and reconnaissance. This work introduces a novel framework that automatically generates alerts and precisely targets specific points of interest for UAV payload delivery using artificial intelligence (AI) algorithms. Our framework seamlessly integrates real-time object detection and recognition capabilities with an advanced AI-driven decision-making module to enable autonomous functionality. The operational process starts with using on board sensors and cameras to continuously monitor the environment. The AI algorithms carefully consider all the options when they identify an object of interest, such as a suspicious package or a predetermined target. They then determine the optimal course of action. If necessary, the system triggers an alert and arranges for a series of actions to enable the UAV to drop a payload at the designated location. Payload delivery mechanisms, advanced object detection algorithms, path planning algorithms that optimize UAV navigation, and deep learning models tailored for object recognition are some of the framework's key elements. The AI-powered decision-making module ensures that choices are made accurately, promptly, and with consideration for the mission's goals, the law, and the environment. A comprehensive evaluation of the proposed framework's efficacy is conducted through simulated and real-world experiments, showcasing its versatility in various contexts. The results Validate the framework's remarkable accuracy in object detection and recognition, as well as its reliable payload delivery capabilities. Furthermore, the system's autonomy lessens the need for human intervention, which makes it particularly suitable for applications that demand accuracy and speed. Our work advances the development of AI-powered unmanned aerial vehicles (UAVs) designed for automated monitoring, law enforcement, and disaster relief operations. With its great potential to improve situational awareness and operational effectiveness in a variety of environments, the proposed framework could usher in a new era of autonomous aerial system advancements.

Experimental Study on Using Synthetic Images as a Portion of Training Dataset for Object Recognition in Construction Site

The application of Artificial Intelligence (AI) across various industries necessitates the acquisition of relevant environmental data and the implementation of AI recognition learning based on this data. However, the data available in real-world environments are limited and difficult to obtain. Construction sites represent dynamic and hazardous environments with a significant workforce, making data acquisition challenging and labor-intensive. To address these issues, this experimental study explored the potential of generating synthetic data to overcome the challenges of obtaining data from hazardous construction sites. Additionally, this research investigated the feasibility of hybrid dataset in securing construction-site data by creating synthetic data for scaffolding, which has a high incidence of falls but low object recognition rates due to its linear object characteristics. We generated a dataset by superimposing scaffolding objects, from which the backgrounds were removed, onto various construction site background images. Using this dataset, we produced a hybrid dataset to assess the feasibility of synthetic data for construction sites and to evaluate improvements in object recognition performance. By finding the optimal composition ratio with real data and conducting model training, the highest accuracy was achieved at an 8:2 ratio, with a construction object recognition accuracy of 0.886. Therefore, this study aims to reduce the risk and labor associated with direct data collection at construction sites through a hybrid dataset, achieving data generation at a low cost and high efficiency. By generating synthetic data to find the optimal ratio and constructing a hybrid dataset, this research demonstrates the potential to address the problems of data scarcity and data quality on construction sites. The improvement in recognition accuracy of the construction safety management system is anticipated, suggesting that the creation of synthetic data for constructing a hybrid dataset can reduce construction safety-accident issues.

CAST: Cross-Modal Retrieval and Visual Conditioning for image captioning

Image captioning requires not only accurate recognition of objects and corresponding relationships, but also full comprehension of the scene information. However, existing models suffer from partial understanding and object hallucination. In this paper, a Cross-modal retrievAl and viSual condiTioning model (CAST) is proposed to address the above issues for image captioning with three key modules: an image–text retriever, an image & memory comprehender and a dual attention decoder. Aiming at a comprehensive understanding, we propose to exploit cross-modal retrieval to mimic human cognition, i.e., to trigger retrieval of contextual information (called episodic memory) about a specific event. Specifically, the image–text retriever searches the top n relevant sentences which serve as episodic memory for each input image. Then the image & memory comprehender encodes an input image and enriches episodic memory by self-attention and relevance attention respectively, which can encourage CAST to comprehend the scene thoroughly and support decoding more effectively. Finally, such image representation and memory are integrated into our dual attention decoder, which performs visual conditioning by re-weighting image and text features to alleviate object hallucination. Extensive experiments are conducted on MS COCO and Flickr30k datasets, which demonstrate that our CAST achieves state-of-the-art performance. Our model also has a promising performance even in low-resource scenarios (i.e. 0.1%, 0.5% and 1% of MS COCO training set).