Accurate breast mass segmentation of automated breast ultrasound (ABUS) images plays a crucial role in 3D breast reconstruction which can assist radiologists in surgery planning. Although the convolutional neural network has great potential for breast mass segmentation due to the remarkable progress of deep learning, the lack of annotated data limits the performance of deep CNNs. In this paper, we present an uncertainty aware temporal ensembling (UATE) model for semi-supervised ABUS mass segmentation. Specifically, a temporal ensembling segmentation (TEs) model is designed to segment breast mass using a few labeled images and a large number of unlabeled images. Considering the network output contains correct predictions and unreliable predictions, equally treating each prediction in pseudo label update and loss calculation may degrade the network performance. To alleviate this problem, the uncertainty map is estimated for each image. Then an adaptive ensembling momentum map and an uncertainty aware unsupervised loss are designed and integrated with TEs model. The effectiveness of the proposed UATE model is mainly verified on an ABUS dataset of 107 patients with 170 volumes, including 8856 2D labeled slices. The Jaccard index (JI), Dice similarity coefficient (DSC), pixel-wise accuracy (AC) and Hausdorff distance (HD) of the proposed method on testing set are 63.65%, 74.25%, 99.21% and 3.81mm respectively. Experimental results demonstrate that our semi-supervised method outperforms the fully supervised method, and get a promising result compared with existing semi-supervised methods.

This validation study of quantitative multiparametric MR image features shows that 10th percentile values, average apparent diffusion coefficient (ADC) values, and T2-weighted signal-intensity skewness are consistently effective across imagers made by two manufacturers for distinguishing prostate cancer from normal-tissue regions of interest, and that the 10th percentile and the average ADC values are significantly, moderately, and negatively correlated with lesion-specific Gleason scores, consistent across imagers made by two manufacturers. PurposeTo validate three previously identified quantitative image features across multiparametric magnetic resonance (MR) images acquired with imagers made by two different manufacturers to differentiate prostate cancer (PC) from normal prostatic tissue and to assess cancer aggressiveness.Materials and MethodsThis study was HIPAA-compliant and approved by the institutional review board. Preoperative 1.5-T multiparametric endorectal MR images of 119 PC patients (dataset A, 71 patients; dataset B, 48 patients) were analyzed, and 265 PC and normal peripheral zone regions of interests (ROIs) were identified through histologic and MR consensus review. The 10th percentile average apparent diffusion coefficient (ADC) value, average ADC value, and skewness of T2-weighted signal-intensity histogram were evaluated with area under the receiver operating characteristic curve (AUC). The image features were combined with a linear discriminant analysis classifier and evaluated both on the image dataset of each type of i...

Infrared and visible image fusion aims to generate one image with comprehensive information. It can maintain rich texture characteristics and thermal information. However, for existing image fusion methods, the fused images either sacrifice the salience of thermal targets and the richness of textures or introduce the interference of useless information like artifacts. To alleviate these problems, an effective cross-modal coordinate attention network for infrared and visible image fusion called CCAFusion is proposed in this paper. To fully integrate complementary features, the cross-modal image fusion strategy based on coordinate attention is designed, which consists of the feature-awareness fusion module and the feature-enhancement fusion module. Moreover, a multiscale skip connection-based network is employed to obtain multiscale features in the infrared image and the visible image, which can fully utilize the multi-level information in the fusion process. To reduce the discrepancy between the fused image and the input images, a multiple constrained loss function including the base loss and the auxiliary loss is developed to adjust the gray-level distribution and ensure the harmonious coexistence of structure and intensity in fused images, thereby preventing the pollution of useless information like artifacts. Extensive experiments conducted on widely used datasets demonstrate that our CCAFusion achieves superior performance over state-of-the-art image fusion methods in both qualitative evaluation and quantitative measurement. Furthermore, the application to salient object detection reveals the potential of our CCAFusion for high-level vision tasks, which can effectively boost the detection performance.

Unsupervised cross domain (UCD) person re-identification (re-ID) aims to apply a model trained on a labeled source domain to an unlabeled target domain. It faces huge challenges as the identities have no overlap between these two domains. At present, most UCD person re-ID methods perform “supervised learning” by assigning pseudo labels to the target domain, which leads to poor re-ID performance due to the pseudo label noise. To address this problem, a multi-loss optimization learning (MLOL) model is proposed for UCD person re-ID. In addition to using the information of clustering pseudo labels from the perspective of supervised learning, two losses are designed from the view of similarity exploration and adversarial learning to optimize the model. Specifically, in order to alleviate the erroneous guidance brought by the clustering error to the model, a ranking-average-based triplet loss learning and a neighbor-consistency-based loss learning are developed. Combining these losses to optimize the model results in a deep exploration of the intra-domain relation within the target domain. The proposed model is evaluated on three popular person re-ID datasets, Market-1501, DukeMTMC-reID, and MSMT17. Experimental results show that our model outperforms the state-of-the-art UCD re-ID methods with a clear advantage.

When dealing with computed tomography volume data, the accurate segmentation of lung nodules is of great importance to lung cancer analysis and diagnosis, being a vital part of computer-aided diagnosis systems. However, due to the variety of lung nodules and the similarity of visual characteristics for nodules and their surroundings, robust segmentation of nodules becomes a challenging problem. A segmentation algorithm based on the fast marching method is proposed that separates the image into regions with similar features, which are then merged by combining regions growing with k-means. An evaluation was performed with two distinct methods (objective and subjective) that were applied on two different datasets, containing simulation data generated for this study and real patient data, respectively. The objective experimental results show that the proposed technique can accurately segment nodules, especially in solid cases, given the mean Dice scores of 0.933 and 0.901 for round and irregular nodules. For non-solid and cavitary nodules the performance dropped—0.799 and 0.614 mean Dice scores, respectively. The proposed method was compared to active contour models and to two modern deep learning networks. It reached better overall accuracy than active contour models, having comparable results to DBResNet but lesser accuracy than 3D-UNet. The results show promise for the proposed method in computer-aided diagnosis applications.

The infrared and visible image fusion aims to generate one image with rich information by integrating thermal regions from the infrared image and texture details from the visible image, which is beneficial to facilitate the capacity of video surveillance and object detection in complex environments. Although there is great progress in image fusion algorithms, artifacts and inconsistencies are still challenging tasks. To alleviate these problems, a multiscale attention network for infrared and visible image fusion (MAFusion) is proposed. The network consists of an encoder, a fusion strategy, and a decoder. Specifically, the encoder is adopted to extract multiscale features by feeding the source images. An attention-based model is then designed as the fusion strategy to integrate different features in the infrared and visible images. The attention-based model can highlight the thermal targets in the infrared image and maintain details in the visible image, so as to avoid the generation of artifacts. The decoder is based on a multiscale skip connection to incorporate low-level details with high-level semantics at different scales. The vital features of infrared and visible images can be fully preserved by the multiscale skip connection network to restrict the introduction of inconsistencies. Furthermore, we develop a feature-preserving loss function to train the proposed network. Experimental results demonstrate that the proposed network delivers advantages and effectiveness compared with the state-of-the-art fusion methods in qualitative and quantitative assessments. Besides, we apply the fused image generated by MAFusion to crowd counting (CC), which can effectively improve the CC performance in low-illumination conditions.