An automatic gastric polyp detection technique using deep learning

Purpose In order to resolve the situation of high missed diagnosis rate and high misdiagnosis rate of the pathological analysis of the gastrointestinal endoscopic images by experts, we propose an automatic polyp detection algorithm based on Single Shot Multibox Detector (SSD). Method In the paper, SSD is based on VGG-16, the fully connected layer is changed to a convolutional layer, and four convolutional layers with successively decreasing scales are added as a new network structure. In order to verify the practicability, it is not only compared with manual polyp detection but also with Mask R-CNN. Results Multiple experimental results show that the mean Average Precision (mAP) of the SSD network is 95.74%, which is 12.4% higher than the manual detection and 5.7% higher than the Mask R-CNN. When detecting a single frame of image, the detection speed of SSD is 8.41 times that of manual detection. Conclusion Based on the traditional pattern recognition algorithm and the target detection algorithm using deep learning, we select a variety of algorithms to identify and classify polyps to achieve efficient detection results. Our research demonstrates that deep learning has a lot of room for development in the field of gastrointestinal image recognition.

Single Shot MultiBox Detector (SSD) is an effective method for multi-targets detection. However, due to the small scale and less information, small targets detection is a difficult assignment for SSD. In order to address this issue, we design an framework based on SSD, namely feature fusion and enhancement for SSD. First, considering that the feature information contained in different scale feature maps is different, especially the shallow feature maps which contains more detailed information and lacks semantic information, we propose a method of fusing different scale feature maps to introduce abundant object information. Secondly, although the feature maps has multiple channels, the importance of different channels' feature to the detected object is different. We introduce the feature enhancement module for effective feature enhancement. Training on the PASCAL VOC2007 and VOC2012, the proposed method has an mAP of 79.07%, which is 1.4% higher than SSD, and an FPS of 39 at the input size of 300 x300.

Accurate identification of vehicles and estimating density in traffic surveillance systems is a challenging task, particularly in scenarios with closely spaced lanes. Single Shot MultiBox Detector (SSD) is introduced in vehicle detection and classification due to its speed and accuracy. It utilizes a transfer learning technique that enables them to utilize features from pretrained Convolutional Neural Networks (CNNs). Although the utilization of multi-scale feature maps in SSD has achieved efficient results in vehicle detection, it struggles to identify key features of vehicles due to its one-stage detection approach. Furthermore, the use of VGG16 as the backbone network in these approaches leads to the loss of fine-grained details, posing challenges in accurately localizing and classifying small vehicles. Also, there is no interaction between high- and low-level features, which restricts the networks ability to effectively integrate and utilize both features for accurate vehicle detection. To overcome these limitations, an improved SSD approach is introduced in this paper, leveraging interactive multi-scale attention characteristics to accurately detect and classify vehicles. This approach utilizes ResNet50 as the network backbone to overcome the limitations of traditional SSDs in detecting small vehicles. Also, an attention block is incorporated into it to focus on key details and assign higher attention to relevant pixels within the feature map. Also, the network employs a parallel detection framework and shares multi-scale layers (both high and low), enabling efficient detection of vehicles with various sizes. Then, traffic density is estimated based on the weights assigned to the categorized vehicles obtained from the improved SSD and the recorded area. Finally, traffic is classified as high, low, or moderate by comparing the estimated density to a threshold value. By adding attention characteristics of different scales to the original detection branch and replacing the VGG16 with ResNet50 of the SSD technique using our method, the feature representation capability and detection accuracy are both significantly improved.

One of the researcher's interests and challenges is animal identification techniques. There are several challenges that researchers in this field confront that limit detection effectiveness and efficiency, such as picture illumination variation, animal occlusion, colour similarity of animal colours with backdrop environment, and so on. The purpose of this study is to detect and classify multi-label images of mammals, which we propose to do using the Single Shot Multi-Box Detector (SSD) and the MobileNet v1 coco_2017 model. Another objective is to locate and identify several items (animals) from the Mammal category in digital photos. Based on deep learning technology, the recommended SSD is regarded as a more accurate, rapid, and efficient technique to recognise objects of various sizes. We utilised 2000 pictures in the network taken from standard datasets (such as Caltech 101) and the net in this proposal. The SSD framework enhances Convolution Neural Network (CNN) detection and identification operations. During the prediction phase, the network assigns scores to the existence of each object class and draws a box around each object in the picture. Each box includes a name that defines the kind of object, and the score denotes the likelihood of the object's association to that category. During the procedure, boxes are changed to get the best fit to the form of the item. The experimental findings of this work demonstrated the efficacy of identifying and detecting animals even when light, position, and occlusion were varied. The detection and classification accuracy can reach 98.7%. Unlike other comparable efforts, this recommendation is more dependable and accurate, and it identifies a wide variety of Mammals species. Keywords— MobileNet , Convolution Neural Network (CNN), Single Shot Multi-Box Detector (SSD)

The use of computer vision and deep learning in boundary documentation for land registration stems from the ongoing demand for appropriate mapping approaches of unregistered land rights to eradicate the global challenge of tenure insecurity. Previous research has yielded promising results towards automated extraction of photo‐visible cadastral boundaries from high‐resolution imagery. Nonetheless, the extraction of invisible cadastral boundaries is still a challenge. This study investigates the place of sensor/s on‐board unmanned aerial vehicles and deep learning algorithms in detecting cadastral boundaries. It develops a participatory boundary marking procedure using low‐cost markers to bring monument to previously invisible and ill‐defined cadastral boundaries. After that, the researchers trained and tested the accuracy of a convolutional neural network, namely single shot multi‐box detector (SSD) based on Residual Neural Network (ResNet) and Visual Geometry Group (VGG) backbone networks to automatically detect cadastral boundary markers from unmanned aerial vehicle imagery. SSD based on ResNet34 performed best with 0.88 precision, 0.92 recall and 0.91 F measure or (F1) score. VGG19‐based SSD yielded a precision of 0.47, recall of 0.53 and F1 score of 0.50. The horizontal accuracy of the cadastral map generated varied from 0.089 to 0.496 m per parcel, with a standard deviation of 0.120 m. Results show that this approach is practical for cadastral mapping in rural areas.

In view of the fact that the feature layers of different scales in the traditional single shot multibox detector (SSD) are independent of each other, resulting in poor detection performance for small objects. We propose an improved SSD network for small object detection based on dilated convolution and feature fusion, which is called DFSSD. Specifically, by introducing dilated convolution, we enhance the receptive field of the third-level feature map in the network, which enables the feature map to obtain more global information. At the same time, we designed feature fusion module to fuse low-level feature map with detailed information and high-level feature map with rich semantic information. We adjust the prediction box scale of the DFSSD network prediction layer. Our proposed network obtains 78.9% mAP on PASCAL VOC2007 test at 40 FPS and 74.7% AP on difficult objects in car class of KITTI dataset. The results outperform the original SSD model by 1.4 and 1.2 points respectively.

Single Shot MultiBox Detector (SSD) method using multi-scale feature maps for object detection, showing outstanding performance in object detection task. However, as a one-stage detection method, it’s difficult for SSD methods to quickly notice significant areas of objects in the image. In the SSD network structure, feature maps of different scales are used to independently predict object, and there is a lack of interaction between low-level feature maps and high-level feature maps. In this paper we propose an enhanced SSD method using interactive multi-scale attention features (MA-SSD). Our method uses the attention mechanism to generate attention features of multiple scales and adds it to the original detection branch of the SSD method, which effectively enhances the feature representation ability and improves the detection accuracy. At the same time, the feature of different detection scales interacts with each other, and all the detection branches in our method have a parallel structure, which ensures the detection efficiency. Our proposed method achieves competitive performance on the public dataset PascalVOC.

Deep learning techniques help computer vision automatically learn the intrinsic patterns within complex data. This research mainly concentrates on creating a mobile application based on augmented reality for elderly mobile users that helps in identifying the traffic signals and other signboards in real‐time using deep learning techniques. TensorFlow serves as an implementation platform to build the object detection system with deep learning. The single shot multibox detector (SSD) model and the two‐stage faster‐regional convolutional neural network (RCNN) models from TensorFlow's object detection application programming interface (API) are compared in this study. The SSD model with MobileNet as a backbone network serves well for this study as it is faster than the RCNN model with comparable accuracy. However, unconstrained environments like occlusions can be an obstacle to the effective performance of an object detection system. This research provides a solution to handle occlusions by developing a robust object detection system through image segmentation techniques. The model introduced is based on the SSD MobileNet model which enables it to be deployable on mobile devices for real‐time offline detection. The developed model exhibits faster performance than the state‐of‐the‐art instance segmentation model, Mask RCNN with comparable accuracy. Elaborated implementation of this system and results are presented in further sections.

Automatic target recognition (ATR) is a challenging task for several computer vision applications. It requires efficient, accurate, and robust methods for target detection and target identification. Deep learning has shown great success in many computer vision applications involving color RGB images. However, the performance of these networks in ATR with infrared sensor data needs further investigation. In this paper, we propose a multistage automatic target detection and recognition (ATDR) system that performs both target detection and target classification on infrared (IR) imagery using deep learning. Our system processes large IR image frames where targets take <1 % of the total number of pixels. First, we train a state-of-the-art object detector you only look once (YOLO) to localize all potential targets in the input image frame. Then, we train a convolutional neural network (CNN) to identify these detections as targets or false alarms. In this second phase, we adapt and analyze the performance of three CNN architectures: a compact and fully connected CNN, VGG16 with batch normalization, and a wide residual neural network (WRN). We also explore the use of a loss function that optimizes directly the area under the receiver operating characteristic (ROC) curve (AUC), and adapt it to our ATR application. To enhance the robustness of the proposed ATR to perturbation and variations introduced during the detection stage, we train our CNN classifiers on automatically detected targets using YOLO, in addition to ground truth bounding boxes and apply selected data augmentation techniques. To simulate real testing environments, where the spatial location of the targets within the image frame is unknown, only YOLO-detected boxes are used during validation. We evaluate our ATDR on a real benchmark dataset that includes different vehicles captured at different resolutions. Our experiments have shown that YOLO can detect most of the targets at the expense of generating a high number of false alarms. We show that the VGG-16 network with batch normalization, which is the best performing model, can correctly identify the classes of the targets, as well as classify the majority of YOLO’s false detections into an additional nontarget class. We also show that the proposed training modification to optimize an AUC-based loss function for ATR proved to be advantageous mainly in identifying difficult targets.

Small bowel polyps exhibit variations related to color, shape, morphology, texture, and size, as well as to the presence of artifacts, irregular polyp borders, and the low illumination condition inside the gastrointestinal GI tract. Recently, researchers developed many highly accurate polyp detection models based on one-stage or two-stage object detector algorithms for wireless capsule endoscopy (WCE) and colonoscopy images. However, their implementation requires a high computational power and memory resources, thus sacrificing speed for an improvement in precision. Although the single-shot multibox detector (SSD) proves its effectiveness in many medical imaging applications, its weak detection ability for small polyp regions persists due to the lack of information complementary between features of low- and high-level layers. The aim is to consecutively reuse feature maps between layers of the original SSD network. In this paper, we propose an innovative SSD model based on a redesigned version of a dense convolutional network (DenseNet) which emphasizes multiscale pyramidal feature maps interdependence called DC-SSDNet (densely connected single-shot multibox detector). The original backbone network VGG-16 of the SSD is replaced with a modified version of DenseNet. The DenseNet-46 front stem is improved to extract highly typical characteristics and contextual information, which improves the model’s feature extraction ability. The DC-SSDNet architecture compresses unnecessary convolution layers of each dense block to reduce the CNN model complexity. Experimental results showed a remarkable improvement in the proposed DC-SSDNet to detect small polyp regions achieving an mAP of 93.96%, F1-score of 90.7%, and requiring less computational time.

The single shot multi-box detector (SSD) exhibits low accuracy in small-object detection; this is because it does not consider the scale contextual information between its layers, and the shallow layers lack adequate semantic information. To improve the accuracy of the original SSD, this paper proposes a new single shot multi-box detector using trident feature and squeeze and extraction feature fusion (SSD-TSEFFM); this detector employs the trident network and the squeeze and excitation feature fusion module. Furthermore, a trident feature module (TFM) is developed, inspired by the trident network, to consider the scale contextual information. The use of this module makes the proposed model robust to scale changes owing to the application of dilated convolution. Further, the squeeze and excitation block feature fusion module (SEFFM) is used to provide more semantic information to the model. The SSD-TSEFFM is compared with the faster regions with convolution neural network features (RCNN) (2015), SSD (2016), and DF-SSD (2020) on the PASCAL VOC 2007 and 2012 datasets. The experimental results demonstrate the high accuracy of the proposed model in small-object detection, in addition to a good overall accuracy. The SSD-TSEFFM achieved 80.4% mAP and 80.2% mAP on the 2007 and 2012 datasets, respectively. This indicates an average improvement of approximately 2% over other models.

The goal of multi-modal learning is to use complementary information on the relevant task provided by the multiple modalities to achieve reliable and robust performance. Recently, deep learning has led significant improvement in multi-modal learning by allowing for fusing high level features obtained at intermediate layers of the deep neural network. This paper addresses a problem of designing robust deep multi-modal learning architecture in the presence of the modalities degraded in quality. We introduce deep fusion architecture for object detection which processes each modality using the separate convolutional neural network (CNN) and constructs the joint feature maps by combining the intermediate features obtained by the CNNs. In order to facilitate the robustness to the degraded modalities, we employ the gated information fusion (GIF) network which weights the contribution from each modality according to the input feature maps to be fused. The combining weights are determined by applying the convolutional layers followed by the sigmoid function to the concatenated intermediate feature maps. The whole network including the CNN backbone and GIF is trained in an end-to-end fashion. Our experiments show that the proposed GIF network offers the additional architectural flexibility to achieve the robust performance in handling some degraded modalities.

Hand gesture recognition has great applications in human-computer interaction (HCI), human-robot interaction (HRI), and supporting the deaf and mute. To build a hand gesture recognition model using deep learning (DL) with high results then needs to be trained on many data and in many different conditions and contexts. In this paper, we publish the TQU-HG dataset of large RGB images with low resolution (640×480) pixels, low light conditions, and fast speed (16 fps). TQU-HG dataset includes 60,000 images collected from 20 people (10 male, 10 female) with 15 gestures of both left and right hands. A comparative study with two branches: i) based on Mediapipe TML and ii) Based on convolutional neural networks (CNNs) (you only look once (YOLO); YOLOv5, YOLOv6, YOLOv7, YOLOv8, YOLO-Nas, single shot multiBox detector (SSD) VGG16, residual network (ResNet)18, ResNext50, ResNet152, ResNext50, MobileNet V3 small, and MobileNet V3 large), the architecture and operation of CNNs models are also introduced in detail. We especially fine-tune the model and evaluate it on TQU-HG and HaGRID datasets. The quantitative results of the training and testing are presented (F1-score of YOLOv8, YOLO-Nas, MobileNet V3 small, ResNet50 is 98.99%, 98.98%, 99.27%, 99.36%, respectively on the TQU-HG dataset and is 99.21%, 99.37%, 99.36%, 86.4%, 98.3%, respectively on the HaGRID dataset). The computation time of YOLOv8 is 6.19 fps on the CPU and 18.28 fps on the GPU.

&lt;p&gt;Colorectal cancer (CRC) is one of the most prevalent malignant tumors in Colombia and the world. These neoplasms originate in adenomatous lesions or polyps that must be resected to prevent the disease, which can be done with a colonoscopy. It has been reported that during colonoscopy polyps are detected in 40% of men and 30% of women (hyperplastic, adenomatous, serrated, among others), and, on average, 25% of adenomatous polyps (main quality indicator in colonoscopy). However, these lesions are not easy to observe due to the multiplicity of blind spots in the colon and the human error associated with the examination. Objective: to create a computational method for the automatic detection of colorectal polyps using artificial intelligence in recorded videos of real colonoscopy procedures. Methodology: public databases with colorectal polyps and a data collection built in a University Hospital were used. Initially, all the frames of the videos were normalized to reduce the high variability between databases. Subsequently, the polyp detection task is done with a deep learning method using a convolutional neural network. This network is initialized with weights learned on millions of national images from the ImageNet database. The weights of the network are updated using colonoscopy images, following the tuning technique. &lt;strong&gt;&lt;/strong&gt;&lt;/p&gt;

To improve the feature recognition ability of deep model transfer learning, we propose a hybrid deep transfer learning method for image classification based on restricted Boltzmann machines (RBM) and convolutional neural networks (CNNs). It integrates learning abilities of two models, which conducts subject classification by exacting structural higher-order statistics features of images. While the method transfers the trained convolutional neural networks to the target datasets, fully-connected layers can be replaced by restricted Boltzmann machine layers; then the restricted Boltzmann machine layers and Softmax classifier are retrained, and BP neural network can be used to fine-tuned the hybrid model. The restricted Boltzmann machine layers has not only fully integrated the whole feature maps, but also learns the statistical features of target datasets in the view of the biggest logarithmic likelihood, thus removing the effects caused by the content differences between datasets. The experimental results show that the proposed method has improved the accuracy of image classification, outperforming other methods on Pascal VOC2007 and Caltech101 datasets.