No abstract

A very large volume of images is uploaded to the Internet daily. However, current hashing methods for image retrieval are designed for static databases only. They fail to consider the fact that the distribution of images can change when new images are added to the database over time. The changes in the distribution of images include both discovery of a new class and a distribution of images within a class owing to concept drift. Retraining of hash tables using all images in the database requires a large computation effort. This is also biased to old data owing to the huge volume of old images which leads to a poor retrieval performance over time. In this paper, we propose the incremental hashing (ICH) method to deal with the two aforementioned types of changes in the data distribution. The ICH uses a multihashing to retain knowledge coming from images arriving over time and a weight-based ranking to make the retrieval results adaptive to the new data environment. Experimental results show that the proposed method is effective in dealing with changes in the database.

This paper addresses a problem in the hashing technique for large scale image retrieval: learn a compact hash code to reduce the storage cost with performance comparable to that of the long hash code. A longer hash code yields a better precision rate of retrieved images. However, it also requires a larger storage, which limits the number of stored images. Current hashing methods employ the same code length for both queries and stored images. We propose a new hashing scheme using two hash codes with different lengths for queries and stored images, i.e., the asymmetric cyclical hashing. A compact hash code is used to reduce the storage requirement, while a long hash code is used for the query image. The image retrieval is performed by computing the Hamming distance of the long hash code of the query and the cyclically concatenated compact hash code of the stored image to yield a high precision and recall rate. Experiments on benchmarking databases consisting up to one million images show the effectiveness of the proposed method. Index Terms-Asymmetric hashing with different code lengths, hashing, large scale image retrieval.1520-9210 and is currently working toward the M.S. degree at the South China University of Technology. His research interests include machine learning, computer vision, information retrieval, and evolutionary computation.

Objectives To investigate the methodology and clinical application of ultrasound attenuation imaging (ATI) and comparative analyze the diagnostic performance of ATI and controlled attenuation parameters (CAP) for detecting and grading hepatic steatosis. Methods A total of 159 patients with NAFLD were prospectively enrolled. CAP and ATI examinations were performed within a week before proton magnetic resonance spectroscopy (1H‐MRS). Ten liver attenuation coefficient (AC) measurements by ATI were obtained in each patient. The interclass correlation coefficients (ICCs) of the intraobserver consistencies and the ICCs between the median of the first two through the first nine measurements and all 10 measurements were calculated. The correlations between 1H‐MRS, CAP, biological data, and ATI were evaluated. The significant factors associated with ATI and the diagnostic performance of ATI and CAP for detecting hepatic steatosis was evaluated. Results The median value of AC for detecting hepatic steatosis was 0.831 dB/cm/MHz. For the intraobserver consistency of ATI, the ICC was 0.931. Compared with 10 measurements, a minimum of four ATI measurements was required. The correlation of AC with hepatic fat fraction (HFF) was significantly higher than that of CAP (0.603 vs 0.326, P = .0015). The HFF and triglyceride (TG) were the significant factors for the ATI. The area under the receiver operating characteristics (ROC) curves of ATI and CAP were 0.939 and 0.788 for detecting ≥10% hepatic steatosis; 0.751 and 0.572 for detecting >33% hepatic steatosis. The cutoff values of ATI and CAP were 0.697 dB/cm/MHz and 310 dB/m for detecting ≥10% hepatic steatosis; 0.793 dB/cm/MHz and 328 dB/m for detecting >33% hepatic steatosis. The sensitivity of ATI and CAP were 85.92% and 52.11% for detecting ≥10% hepatic steatosis; 87.50% and 82.14% for detecting >33% hepatic steatosis. The specificity of ATI and CAP were 94.12% and 100% for detecting ≥10% hepatic steatosis; 54.37% and 43.69% for detecting >33% hepatic steatosis. Conclusions ATI technology showed excellent intraobserver consistency and the optimal minimum number of ATI measurements was 4. ATI is a promising noninvasive, quantitative and convenient tool for assessing hepatic steatosis.