Locating Target Regions for Image Retrieval in an Unsupervised Manner.

Image retrieval performance can be improved by training a convolutional neural network (CNN) model with annotated data to facilitate accurate localization of target regions. However, obtaining sufficiently annotated data is expensive and impractical in real settings. It is challenging to achieve accurate localization of target regions in an unsupervised manner. To address this problem, we propose a new unsupervised image retrieval method named unsupervised target region localization (UTRL) descriptors. It can precisely locate target regions without supervisory information or learning. Our method contains three highlights: 1) we propose a novel zero-label transfer learning method to address the problem of co-localization in target regions. This enhances the potential localization ability of pretrained CNN models through a zero-label data-driven approach; 2) we propose a multiscale attention accumulation method to accurately extract distinguishable target features. It distinguishes the importance of features by using local Gaussian weights; and 3) we propose a simple yet effective method to reduce vector dimensionality, named twice-PCA-whitening (TPW), which reduces the performance degradation caused by feature compression. Notably, TPW is a robust and general method that can be widely applied to image retrieval tasks to improve retrieval performance. This work also facilitates the development of image retrieval based on short vector features. Extensive experiments on six popular benchmark datasets demonstrate that our method achieves about 7% greater mean average precision (mAP) compared to existing state-of-the-art unsupervised methods.

Diagnosis of Alzheimer's Disease Based on Deeply-Fused Nets.

AIM AND OBJECTIVE: Fast and accurate diagnosis of Alzheimer's disease is very important for the care and further treatment of patients. Along with the development of deep learning, impressive progress has also been made in the automatic diagnosis of AD. Most existing studies on automatic diagnosis are concerned with a single base network, whose accuracy for disease diagnosis still needs to be improved. This study was undertaken to propose a method to improve the accuracy of the automatic diagnosis of AD. MATERIALS AND METHODS: MRI image data from the Alzheimer's Disease Neuroimaging Initiative were used to train a deep learning model to achieve a computer-aided diagnosis of Alzheimer's disease. The data consisted of 138 with AD, 280 with mild cognitive impairment, and 138 normal controls. Here, a new deeply-fused net is proposed, which combines several deep convolutional neural networks, thereby avoiding the error of a single base network and increasing the classification accuracy and generalization capacity. RESULTS: Experiments show that when differentiating between patients with AD, mild cognitive impairment, and normal controls on a subset of the ADNI database without data leakage, the new architecture improves the accuracy by about 4 percentage points as compared to a single standard based network. CONCLUSION: This new approach exhibits better performance, but there is still much to be done before its clinical application. In the future, greater research effort will be devoted to improving the performance of the deeply-fused net.