An Optimal Artificial Intelligence System for Real-Time Endoscopic Prediction of Invasion Depth in Early Gastric Cancer.

We previously constructed a VGG-16 based artificial intelligence (AI) model (image classifier [IC]) to predict the invasion depth in early gastric cancer (EGC) using endoscopic static images. However, images cannot capture the spatio-temporal information available during real-time endoscopy-the AI trained on static images could not estimate invasion depth accurately and reliably. Thus, we constructed a video classifier [VC] using videos for real-time depth prediction in EGC. We built a VC by attaching sequential layers to the last convolutional layer of IC v2, using video clips. We computed the standard deviation (SD) of output probabilities for a video clip and the sensitivities in the manner of frame units to observe consistency. The sensitivity, specificity, and accuracy of IC v2 for static images were 82.5%, 82.9%, and 82.7%, respectively. However, for video clips, the sensitivity, specificity, and accuracy of IC v2 were 33.6%, 85.5%, and 56.6%, respectively. The VC performed better analysis of the videos, with a sensitivity of 82.3%, a specificity of 85.8%, and an accuracy of 83.7%. Furthermore, the mean SD was lower for the VC than IC v2 (0.096 vs. 0.289). The AI model developed utilizing videos can predict invasion depth in EGC more precisely and consistently than image-trained models, and is more appropriate for real-world situations.

Malignant thoracic lymph node classification with deep convolutional neural networks on real-time endobronchial ultrasound (EBUS) images.

BACKGROUND: Thoracic lymph node (LN) evaluation is essential for the accurate diagnosis of lung cancer and deciding the appropriate course of treatment. Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is considered a standard method for mediastinal nodal staging. This study aims to build a deep convolutional neural network (CNN) for the automatic classification of metastatic malignancies involving thoracic LN, using EBUS-TBNA. METHODS: Patients who underwent EBUS-TBNAs to assess the presence of malignancy in mediastinal LNs during a ten-month period at Severance Hospital, Seoul, Republic of Korea, were included in the study. Corresponding LN ultrasound images, pathology reports, demographic data, and clinical history were collected and analyzed. RESULTS: A total of 2,394 endobronchial ultrasound (EBUS) images of 1,459 benign LNs from 193 patients, and 935 malignant LNs from 177 patients, were collected. We employed the visual geometry group (VGG)-16 network to classify malignant LNs using only traditional cross-entropy for classification loss. The sensitivity, specificity, and accuracy of predicting malignancy were 69.7%, 74.3%, and 72.0%, respectively, and the overall area under the curve (AUC) was 0.782. We applied the new loss function to train the network and, using the modified VGG-16, the AUC improved to a value of 0.8. The sensitivity, specificity, and accuracy improved to 72.7%, 79.0%, and 75.8%, respectively. In addition, the proposed network can process 63 images per second on a single mainstream graphics processing unit (GPU) device, making it suitable for real-time analysis of EBUS images. CONCLUSIONS: Deep CNNs can effectively classify malignant LNs from EBUS images. Selecting LNs that require biopsy using real-time EBUS image analysis with deep learning is expected to shorten the EBUS-TBNA procedure time, increase lung cancer nodal staging accuracy, and improve patient safety.

Deep learning for early dental caries detection in bitewing radiographs.

The early detection of initial dental caries enables preventive treatment, and bitewing radiography is a good diagnostic tool for posterior initial caries. In medical imaging, the utilization of deep learning with convolutional neural networks (CNNs) to process various types of images has been actively researched, with promising performance. In this study, we developed a CNN model using a U-shaped deep CNN (U-Net) for caries detection on bitewing radiographs and investigated whether this model can improve clinicians' performance. The research complied with relevant ethical regulations. In total, 304 bitewing radiographs were used to train the CNN model and 50 radiographs for performance evaluation. The diagnostic performance of the CNN model on the total test dataset was as follows: precision, 63.29%; recall, 65.02%; and F1-score, 64.14%, showing quite accurate performance. When three dentists detected caries using the results of the CNN model as reference data, the overall diagnostic performance of all three clinicians significantly improved, as shown by an increased sensitivity ratio (D1, 85.34%; D1', 92.15%; D2, 85.86%; D2', 93.72%; D3, 69.11%; D3', 79.06%; p < 0.05). These increases were especially significant (p < 0.05) in the initial and moderate caries subgroups. The deep learning model may help clinicians to diagnose dental caries more accurately.

A Lesion-Based Convolutional Neural Network Improves Endoscopic Detection and Depth Prediction of Early Gastric Cancer.

In early gastric cancer (EGC), tumor invasion depth is an important factor for determining the treatment method. However, as endoscopic ultrasonography has limitations when measuring the exact depth in a clinical setting as endoscopists often depend on gross findings and personal experience. The present study aimed to develop a model optimized for EGC detection and depth prediction, and we investigated factors affecting artificial intelligence (AI) diagnosis. We employed a visual geometry group(VGG)-16 model for the classification of endoscopic images as EGC (T1a or T1b) or non-EGC. To induce the model to activate EGC regions during training, we proposed a novel loss function that simultaneously measured classification and localization errors. We experimented with 11,539 endoscopic images (896 T1a-EGC, 809 T1b-EGC, and 9834 non-EGC). The areas under the curves of receiver operating characteristic curves for EGC detection and depth prediction were 0.981 and 0.851, respectively. Among the factors affecting AI prediction of tumor depth, only histologic differentiation was significantly associated, where undifferentiated-type histology exhibited a lower AI accuracy. Thus, the lesion-based model is an appropriate training method for AI in EGC. However, further improvements and validation are required, especially for undifferentiated-type histology.

Multiple Objects Fusion Tracker Using a Matching Network for Adaptively Represented Instance Pairs.

Multiple-object tracking is affected by various sources of distortion, such as occlusion, illumination variations and motion changes. Overcoming these distortions by tracking on RGB frames, such as shifting, has limitations because of material distortions caused by RGB frames. To overcome these distortions, we propose a multiple-object fusion tracker (MOFT), which uses a combination of 3D point clouds and corresponding RGB frames. The MOFT uses a matching function initialized on large-scale external sequences to determine which candidates in the current frame match with the target object in the previous frame. After conducting tracking on a few frames, the initialized matching function is fine-tuned according to the appearance models of target objects. The fine-tuning process of the matching function is constructed as a structured form with diverse matching function branches. In general multiple object tracking situations, scale variations for a scene occur depending on the distance between the target objects and the sensors. If the target objects in various scales are equally represented with the same strategy, information losses will occur for any representation of the target objects. In this paper, the output map of the convolutional layer obtained from a pre-trained convolutional neural network is used to adaptively represent instances without information loss. In addition, MOFT fuses the tracking results obtained from each modality at the decision level to compensate the tracking failures of each modality using basic belief assignment, rather than fusing modalities by selectively using the features of each modality. Experimental results indicate that the proposed tracker provides state-of-the-art performance considering multiple objects tracking (MOT) and KITTIbenchmarks.

Object Detection and Classification by Decision-Level Fusion for Intelligent Vehicle Systems.

To understand driving environments effectively, it is important to achieve accurate detection and classification of objects detected by sensor-based intelligent vehicle systems, which are significantly important tasks. Object detection is performed for the localization of objects, whereas object classification recognizes object classes from detected object regions. For accurate object detection and classification, fusing multiple sensor information into a key component of the representation and perception processes is necessary. In this paper, we propose a new object-detection and classification method using decision-level fusion. We fuse the classification outputs from independent unary classifiers, such as 3D point clouds and image data using a convolutional neural network (CNN). The unary classifiers for the two sensors are the CNN with five layers, which use more than two pre-trained convolutional layers to consider local to global features as data representation. To represent data using convolutional layers, we apply region of interest (ROI) pooling to the outputs of each layer on the object candidate regions generated using object proposal generation to realize color flattening and semantic grouping for charge-coupled device and Light Detection And Ranging (LiDAR) sensors. We evaluate our proposed method on a KITTI benchmark dataset to detect and classify three object classes: cars, pedestrians and cyclists. The evaluation results show that the proposed method achieves better performance than the previous methods. Our proposed method extracted approximately 500 proposals on a 1226 × 370 image, whereas the original selective search method extracted approximately 10 6 × n proposals. We obtained classification performance with 77.72% mean average precision over the entirety of the classes in the moderate detection level of the KITTI benchmark dataset.