Recent progress in deep learning, especially deep convolutional neural networks (DCNNs), has led to significant improvement in natural image classification. However, research is still ongoing in the domain of medical image analysis in part due to the shortage of annotated data sets for training DCNNs, the imbalanced number of positive and negative samples, and the difference between medical images and natural images. In this paper, two strategies are proposed to train a DCNN for pathological lymph node image classification. Firstly, the transfer learning strategy is used to deal with the shortage of training samples. Second, a novel exponential loss function is presented for the imbalance in training samples. Four state-of-the-art DCNNs (GoogleNet, ResNet101, Xception, and MobileNetv2) are tested. The experiments demonstrate that the two strategies are effective to improve the performance of pathological lymph node image classification in terms of accuracy and sensitivity with a mean of 0.13% and 1.50%, respectively, for the four DCNNs. In particular, the proposed exponential loss function improved the sensitivity by 3.9% and 4.0% for Xception and ResNet101, respectively.
The recently developed body-wide Automatic Anatomy Recognition (AAR) methodology depends on fuzzy modeling of individual objects, hierarchically arranging objects, constructing an anatomy ensemble of these models, and a dichotomous object recognition–delineation process. The parent-to-offspring spatial relationship in the object hierarchy is crucial in the AAR method. We have found this relationship to be quite complex, and as such any improvement in capturing this relationship information in the anatomy model will improve the process of recognition itself. Currently, the method encodes this relationship based on the layout of the geometric centers of the objects. Motivated by the concept of virtual landmarks (VLs), this paper presents a new one-shot AAR recognition method that utilizes the VLs to learn object relationships by training a neural network to predict the pose and the VLs of an offspring object given the VLs of the parent object in the hierarchy. We set up two neural networks for each parent-offspring object pair in a body region, one for predicting the VLs and another for predicting the pose parameters. The VL-based learning/prediction method is evaluated on two object hierarchies involving 14 objects. We utilize 54 computed tomography (CT) image data sets of head and neck cancer patients and the associated object contours drawn by dosimetrists for routine radiation therapy treatment planning. The VL neural network method is found to yield more accurate object localization than the currently used simple AAR method.
Currently, there are many papers that have been published on the detection and segmentation of lymph nodes from medical images. However, it is still a challenging problem owing to low contrast with surrounding soft tissues and the variations of lymph node size and shape on computed tomography (CT) images. This is particularly very difficult on low-dose CT of PET/CT acquisitions. In this study, we utilize our previous automatic anatomy recognition (AAR) framework to recognize the thoracic-lymph node stations defined by the International Association for the Study of Lung Cancer (IASLC) lymph node map. The lymph node stations themselves are viewed as anatomic objects and are localized by using a one-shot method in the AAR framework. Two strategies have been taken in this paper for integration into AAR framework. The first is to combine some lymph node stations into composite lymph node stations according to their geometrical nearness. The other is to find the optimal parent (organ or union of organs) as an anchor for each lymph node station based on the recognition error and thereby find an overall optimal hierarchy to arrange anchor organs and lymph node stations. Based on 28 contrast-enhanced thoracic CT image data sets for model building, 12 independent data sets for testing, our results show that thoracic lymph node stations can be localized within 2-3 voxels compared to the ground truth.
Face image symmetry is an important factor affecting the accuracy of automatic face recognition. Selecting high symmetrical face image could improve the performance of the recognition. In this paper, we proposed a novel facial symmetry evaluation scheme based on geometric features, including centroid, singular value, in-plane rotation angle of face and the structural similarity index (SSIM). First, we calculate the value of the four features according to the corresponding formula. Then, we use fuzzy logic algorithm to integrate the value of the four features into a single number which represents the facial symmetry. The proposed method is efficient and can adapt to different recognition methods. Experimental results demonstrate its effectiveness in improving the robustness of face detection and recognition.